Appendix — United States v. American Library Assn., Inc.

‘Rise 5 2

No.

In the Supreme Court of the United States

UNITED STATES OF AMERICA, ET AL., APPELLANTS

V.

AMERICAN LIBRARY ASSOCIATION, ET AL.

ON APPEAL FROM THE UNITED STATES DISTRICT COURT

FOR THE EASTERN DISTRICT OF PENNSYLVANIA

APPENDIX TO THE

JURISDICTIONAL STATEMENT

THEODORE B. OLSON

Solicitor General

Counsel of Record

ROBERT D. MCCALLUM, JR.

Assistant Attorney General

EDWIN S. KNEEDLER

Deputy Solicitor General

IRVING L. GORNSTEIN

Assistant to the Solicitor

General

BARBARA L. HERWIG

JACOB M. LEWIS

H. THOMAS BYRON III

AUGUST E. FLENTJE

Attorneys

rtment of Justice

Washington, D.C. 20530-0001

(202) 514-2217

7 — ——E

—————. T

TABLE OF CONTENTS

Page

Appendix A — 1a

1K—!ꝛ—KK 192a

P| + 209a

(I)

APPENDIX A

UNITED STATES DISTRICT COURT

FOR THE EASTERN DISTRICT OF PENNSYLVANIA

Nos. CIV. A. 01-1303 CIV. A. 01-1322

AMERICAN LIBRARY ASSOCIATION, INC., ET AL.

V.

UNITED STATES, ET AL.

MULTNOMAH COUNTY PUBLIC LIBRARY, ET AL.

V.

UNITED STATES OF AMERICA, ET AL.

[May 31, 2002]

OPINION OF THE COURT

Before: BECKER, Chief Circuit Judge, FULLAM and

BARTLE, District Judges.

EDWARD R. BECKER, Chief Circuit Judge.

I. Preliminary Statement

This case challenges an act of Congress that makes

the use of filtering software by public libraries a condi-

tion of the receipt of federal funding. The Internet, as

is well known, is a vast, interactive medium based on a

decentralized network of computers around the world.

Its most familiar feature is the World Wide Web (the

“Web”), a network of computers known as servers that

provide content to users. The Internet provides easy

(la)

2a

access to anyone who wishes to provide or distribute

information to a worldwide audience; it is used by more

than 143 million Americans. Indeed, much of the

world’s knowledge accumulated over centuries is avail-

able to Internet users almost instantly. Approximately

10% of the Americans who use the Internet access it at

public libraries. And approximately 95% of all public

libraries in the United States provide public access to

the Internet.

While the beneficial effect of the Internet in ex-

panding the amount of information available to its users

is self-evident, its low entry barriers have also led to a

perverse result—facilitation of the widespread dis-

semination of hardcore pornography within the easy

reach not only of adults who have every right to access

it (so long as it is not legally obscene or child porno-

graphy), but also of children and adolescents to whom it

may be quite harmful. The volume of pornography on

the Internet is huge, and the record before us demon-

strates that public library patrons of all ages, many

from ages 11 to 15, have regularly sought to access it in

public library settings. There are more than 100,000

pornographic Web sites that can be accessed for free

and without providing any registration information, and

tens of thousands of Web sites contain child porno-

graphy.

Libraries have reacted to this situation by utilizing a

number of means designed to insure that patrons avoid

illegal (and unwanted) content while also enabling

patrons to find the content they desire. Some libraries

have trained patrons in how to use the Internet while

avoiding illegal content, or have directed their patrons

to “preferred” Web sites that librarians have reviewed.

Other libraries have utilized such devices as recessing

3a

the computer monitors, installing privacy screens, and

monitoring implemented by a “tap on the shoulder” of

patrons perceived to be offending library policy. Still

others, viewing the foregoing approaches as inadequate

or uncomfortable (some librarians do not wish to con-

front patrons), have purchased commercially available

software that blocks certain categories of material

deemed by the library board as unsuitable for use in

their facilities. Indeed, 7% of American public libraries

use blocking software for adults. Although such pro-

grams are somewhat effective in blocking large quanti-

ties of pornography, they are blunt instruments that

not only “underblock,” i.e., fail to block access to sub-

stantial amounts of content that the library boards wish

to exclude, but also, central to this litigation, “over-

block,” i.e., block access to large quantities of material

that library boards do not wish to exclude and that is

constitutionally protected.

Most of the libraries that use filtering software seek

to block sexually explicit speech. While most libraries

include in their physical collection copies of volumes

such as The Joy of Sex and The Joy of Gay Sex, which

contain quite explicit photographs and descriptions,

filtering software blocks large quantities of other,

adults and teenagers seek on the Web. One teenager

testified that the Internet access in a public library was

the only venue in which she could obtain information

important to her about her own sexuality. Another

library patron witness described using the Internet to

research breast cancer and reconstructive surgery for

his mother who had breast surgery. Even though some

filtering programs contain exceptions for health and

education, the exceptions do not solve the problem of

4a

overblocking constitutionally protected material. More-

over, as we explain below, the filtering software on

which the parties presented evidence in this case

overblocks not only information relating to health and

sexuality that might be mistaken for pornography or

erotica, but also vast numbers of Web pages and sites

that could not even arguably be construed as harmful or

inappropriate for adults or minors.

The Congress, sharing the concerns of many library

boards, enacted the Children’s Internet Protection Act

(“CIPA”), Pub. L. No. 106-554, which makes the use of

filters by a public library a condition of its receipt of

two kinds of subsidies that are important (or even

critical) to the budgets of many public libraries—grants

under the Library Services and Technology Act, 20

U.S.C. § 9101 et seg. (“LSTA”), and so-called “E-rate

discounts” for Internet access and support under the

Telecommunications Act, 47 U.S.C. § 254. LSTA grant

funds are awarded, inter alia, in order to: (1) assist

libraries in accessing information through electronic

networks, and (2) provide targeted library and informa-

tion services to persons having difficulty using a library

and to underserved and rural communities, including

children from families with incomes below the poverty

line. E-rate discounts serve the similar purpose of

extending Internet access to schools and libraries in

low- income communities. CIPA requires that libraries,

in order to receive LSTA funds or E-rate discounts,

certify that they are using a “technology protection

measure” that prevents patrons from accessing “visual

depictions” that are “obscene,” “child pornography,” or

in the case of minors, “harmful to minors.” 20 U.S.C.

— (LSTA); 47 U.S.C. § 254(h)(6)(B) & (C)

rate).

5a

The plaintiffs, a group of libraries, library associa-

tions, library patrons, and Web site publishers, brought

this suit against the United States and others alleging

that CIPA is facially unconstitutional because: (1) it

induces public libraries to violate their patrons’ First

Amendment rights contrary to the requirements of

South Dakota v. Dole, 483 U.S. 203, 107 S. Ct. 2793, 97

L. Ed. 2d 171 (1987); and (2) it requires libraries to

relinquish their First Amendment rights as a condition

on the receipt of federal funds and is therefore imper-

missible under the doctrine of unconstitutional condi-

tions. In arguing that CIPA will induce public libraries

to violate the First Amendment, the plaintiffs contend

that given the limits of the filtering technology, CIPA’s

conditions effectively require libraries to impose

content-based restrictions on their patrons’ access to

constitutionally protected speech. According to the

plaintiffs, these content-based restrictions are subject

to strict scrutiny under public forum doctrine, see

v. Rector & Visitors of Univ. of Va., 515

US. 819, 837, 115 S. Ct. 2510, 132 L. Ed. 2d 700 (1995),

and are therefore permissible only if they are narrowly

tailored to further a compelling state interest and no

less restrictive alternatives would further that interest,

see Reno v. ACLU, 521 U.S. 844, 874, 117 S. Ct. 2329,

138 L. Ed. 2d 874 (1997). The government responds

1 Plaintiffs advance three other alternative, independent

grounds for holding CIPA facially invalid. First, they submit that

even if CIPA will not induce public libraries to violate the First

Amendment, CIPA nonetheless imposes an unconstitutional condi-

tion on public libraries by requiring them to relinquish their own

First Amendment rights to provide unfiltered Internet access as a

condition on their receipt of federal funds. See infra n. 36. Second,

plaintiffs contend that CIPA is facially invalid because it effects an

impermissible prior restraint on speech by granting filtering

6a

that CIPA will not induce public libraries to violate the

First Amendment, since it is possible for at least some

public libraries to constitutionally comply with CIPA’s

conditions. Even if some libraries’ use of filters might

violate the First Amendment, the government submits

that CIPA can be facially invalidated only if it is im-

possible for any public library to comply with its condi-

tions without violating the First Amendment.

Pursuant to CIPA, a three-judge Court was con-

vened to try the issues. Pub. L. No. 106-554. Following

an intensive period of discovery on an expedited

schedule to allow public libraries to know whether they

need to certify compliance with CIPA by July 1, 2002,

to receive subsidies for the upcoming year, the Court

conducted an eight-day trial at which we heard 20

witnesses, and received numerous depositions, stipula-

tions and documents. The principal focus of the trial

was on the capacity of currently available filtering soft-

ware. The plaintiffs adduced substantial evidence not

only that filtering programs bar access to a substantial

amount of speech on the Internet that is clearly con-

stitutionally protected for adults and minors, but also

that these programs are intrinsically unable to block

only illegal Internet content while simultaneously al-

lowing access to all protected speech.

companies and library staff unfettered discretion to suppress

speech before it has been received by library patrons and before it

has been subject to a judicial determination that it is

under the First Amendment. See Southeastern Promotions, Ltd.

v. Conrad, 420 U.S. 546, 558, 95 S. Ct. 1239, 48 L. Ed. 2d 448 (1975).

Finally, plaintiffs submit that CIPA is unconstitutionally vague.

' See City of Chicago v. Morales, 527 U.S. 41, 119 S. Ct. 1849, 144 L.

Ed. 2d 67 (1999).

7a

As our extensive findings of fact refleet, the plaintiffs

demonstrated that thousands of Web pages containing

protected speech are wrongly blocked by the four

leading filtering programs, and these pages represent

only a fraction of Web pages wrongly blocked by the

programs. The plaintiffs’ evidence explained that the

problems faced by the manufacturers and vendors of

filtering software are legion. The Web is extremely

dynamic, with an estimated 1.5 million new pages added

every day and the contents of existing Web pages

changing very rapidly. The category lists maintained

by the blocking programs are considered to be proprie-

tary information, and hence are unavailable to cus-

tomers or the general public for review, so that public

libraries that select categories when implementing

filtering software do not really know what they are

blocking.

There are many reasons why filtering software

suffers from extensive over-and underblocking, which

we will explain below in great detail. They center on

the limitations on filtering companies’ ability to:

(1) accurately collect Web pages that potentially fall

into a blocked category (e.g., pornography); (2) review

and categorize Web pages that they have collected; and

(3) engage in regular re-review of Web pages that they

have previously reviewed. These failures spring from

constraints on the technology of automated classifica-

tion systems, and the limitations inherent in human

review, including error, misjudgment, and scarce

resources, which we describe in detail infra at 58-74.

One failure of critical importance is that the automated

systems that filtering companies use to collect Web

pages for classification are able to search only text, not

images. This is crippling to filtering companies’ ability

8a

to collect pages containing “visual depictions” that are

obscene, child pornography, or harmful to minors, as

CIPA requires. As will appear, we find that it is cur-

rently impossible, given the Internet’s size, rate of

growth, rate of change, and architecture, and given the

state of the art of automated classification systems, to

develop a filter that neither underblocks nor overblocks

a substantial amount of speech.

The government, while acknowledging that the

filtering software is imperfect, maintains that it is none-

theless quite effective, and that it successfully blocks

the vast majority of the Web pages that meet filtering

companies’ category definitions (e.g., pornography).

The government contends that no more is required. In

its view, so long as the filtering software selected by

the libraries screens out the bulk of the Web pages

proscribed by CIPA, the libraries have made a reason-

able choice which suffices, under the applicable legal

principles, to pass constitutional muster in the context

of a facial challenge. Central to the government’s posi-

tion is the analogy it advances between Internet filter-

ing and the initial decision of a library to determine

which materials to purchase for its print collection.

Public libraries have finite budgets and must make

choices as to whether to purchase, for example, books

on gardening or books on golf. Such content-based

decisions, even the plaintiffs concede, are subject to

rational basis review and not a stricter form of First

Amendment scrutiny. In the government’s view, the

fact that the Internet reverses the acquisition process

and requires the libraries to, in effect, purchase the

entire Internet, some of which (e.g., hardcore perno-

graphy) it does not want, should not mean that it is

9a

chargeable with censorship when it filters out offending

material.

The legal context in which this extensive factual

record is set is complex, implicating a number of consti-

tutional doctrines, including the constitutional limita-

tions on Congress’s spending clause power, the uncon-

stitutional conditions doctrine, and subsidiary to these

issues, the First Amendment doctrines of prior re-

straint, vagueness, and overbreadth. There are a num-

ber of potential entry points into the analysis, but the

most logical is the spending clause jurisprudence in

which the seminal case is South Dakota v. Dole, 483

U.S. 203, 107 S. Ct. 2793, 97 L. Ed. 2d 171 (1987). Dole

outlines four categories of constraints on Congress's

exercise of its power under the Spending Clause, but

the only Dole condition disputed here is the fourth and

last, i.e., whether CIPA requires libraries that receive

LSTA funds or E-rate discounts to violate the con-

stitutional rights of their patrons. As will appear, the

question is not a simple one, and turns on the level of

scrutiny applicable to a public library’s content-based

restrictions on patrons’ Internet access. Whether such

restrictions are subject to strict scrutiny, as plaintiffs

contend, or only rational basis review, as the govern-

ment contends, depends on public forum doctrine.

The government argues that, in providing Internet

access, public libraries do not create a public forum,

since public libraries may reserve the right to exclude

certain speakers from availing themselves of the forum.

Accordingly, the government contends that public

libraries’ restrictions on patrons’ Internet access are

subject only to rational basis review.

Plaintiffs respond that the government's ability to

restrict speech on its own property, as in the case of

10a

restrictions on Internet access in public libraries, is not

unlimited, and that the more widely the state facilitates

the dissemination of private speech in a given forum,

the more vulnerable the state’s decision is to restrict

access to speech in that forum. We agree with the

plaintiffs that public libraries’ content-based restric-

tions on their patrons’ Internet access are subject to

strict scrutiny. In providing even filtered Internet

access, public libraries create a public forum open to

any speaker around the world to communicate with

library patrons via the Internet on a virtually unlimited

number of topics. Where the state provides access to a

“vast democratic forum[ ],” Reno v. ACLU, 521 U.S.

844, 868, 117 S. Ct. 2329, 138 L. Ed. 2d 874 (1997), open

to any member of the public to speak on subjects “as

diverse as human thought,” id. at 870, 117 S. Ct. 2329

(internal quotation marks and citation omitted), the

state’s decision selectively to exclude from the forum

speech whose content the state disfavors is subject to

strict scrutiny, as such exclusions risk distorting the

marketplace of ideas that the state has facilitated.

Application of strict scrutiny finds further support in

the extent to which public libraries’ provision of Inter-

net access uniquely promotes First Amendment values

in a manner analogous to traditional public fora such as

streets, sidewalks, and parks, in which content-based

restrictions are always subject to strict scrutiny.

Under strict scrutiny, a public library’s use of

filtering software is permissible only if it is narrowly

tailored to further a compelling government interest

and no less restrictive alternative would serve that

interest. We acknowledge that use of filtering software

furthers public libraries’ legitimate interests in pre-

. venting patrons from accessing visual depictions of

lla

obscenity, child pornography, or in the case of minors,

material harmful to minors. Moreover, use of filters

also helps prevent patrons from being unwillingly

exposed to patently offensive, sexually explicit content

on the Internet.

We are sympathetic to the position of the gov-

ernment, believing that it would be desirable if there

were a means to ensure that public library patrons

could share in the informational bonanza of the Internet

while being insulated from materials that meet CIPA’s

definitions, that is, visual depictions that dre obscene,

child pornography, or in the case of minors, harmful to

minors. Unfortunately this outcome, devoutly to be

wished, is not available in this less than best of all

possible worlds. No category definition used by the

blocking programs is identical to the legal definitions of

obscenity, child pornography, or material harmful to

minors, and, at all events, filtering programs fail to

block access to a substantial amount of content on the

Internet that falls into the categories defined by CIPA.

As will appear, we credit the testimony of plaintiffs’ ex-

pert Dr. Geoffrey Nunberg that the blocking software

is (at least for the foreseeable future) incapable of

effectively blocking the majority of materials in the

categories defined by CIPA without overblocking a

substantial amount of materials. Nunberg’s analysis

was supported by extensive record evidence. As noted

above. this inability to prevent both substantial

amounts of underblocking and overblocking stems from

several sources, including limitations on the technology

that software filtering companies use to gather and

review Web pages, limitations on resources for human

review of Web pages, and the necessary error that

results from human review processes.

12a

Because the filtering software mandated by CIPA

will block access to substantial amounts of constitu-

tionally protected speech whose suppression serves no

legitimate government interest, we are persuaded that

a public library’s use of software filters is not narrowly

tailored to further any of these interests. Moreover,

less restrictive alternatives exist that further the gov-

ernment’s legitimate interest in preventing the dis-

semination of obscenity, child pornography, and mate-

rial harmful to minors, and in preventing patrons from

being unwillingly exposed to patently offensive, sexu-

ally explicit content. To prevent patrons from access-

ing visual depictions that are obscene and child pornog-

raphy, public libraries may enforce Internet use policies

that make clear to patrons that the library’s Internet

terminals may not be used to access illegal speech.

Libraries may then impose penalties on patrons who

violate these policies, ranging from a warning to noti-

fication of law enforcement, in the appropriate case.

Less restrictive alternatives to filtering that further

libraries’ interest in preventing minors from exposure

to visual depictions that are harmful to minors include

requiring parental consent to or presence during un-

filtered access, or restricting minors’ unfiltered access

to terminals within view of library staff. Finally, op-

tional filtering, privacy screens, recessed monitors, and

placement of unfiltered Internet terminals outside of

sight-lines provide less restrictive alternatives for li-

braries to prevent patrons from being unwillingly ex-

posed to sexually explicit content on the Internet.

In an effort to avoid the potentially fatal legal impli-

cations of the overblocking problem, the government

falls back on the ability of the libraries, under CIPA’s

- disabling provisions, see CIPA §1712 (codified at 20

13a

U.S.C. § 9134(f)(3)), CIPA § 1721(b) (codified at 47

U.S.C. § 254(h)(6)(D)), to unblock a site that is patently

proper yet improperly blocked. The evidence reflects

that libraries can and do unblock the filters when a

patron so requests. But it also reflects that requiring

library patrons to ask for a Web site to be unblocked

will deter many patrons because they are embarrassed,

or desire to protect their privacy or remain anonymous.

Moreover, the unblocking may take days, and may be

unavailable, especially in branch libraries, which are

often less well staffed than main libraries. Accordingly,

CIPA’s disabling provisions do not cure the con-

stitutional deficiencies in public libraries’ use of Inter-

net filters.

Under these circumstances we are constrained to

conclude that the library plaintiffs must prevail in their

contention that CIPA requires them to violate the First

Amendment rights of their patrons, and accordingly is

facially invalid, even under the standard urged on us by

the government, which would permit us to facially

invalidate CIPA only if it is impossible for a single

public library to comply with CIPA’s conditions without

violating the First Amendment. In view of the limita-

tions inherent in the filtering technology mandated by

CIPA, any public library that adheres to CIPA’s condi-

tions will necessarily restrict patrons’ access to a

substantial amount of protected speech, in violation of

the First Amendment. Given this conclusion, we need

not reach plaintiffs’ arguments that CIPA effects a

prior restraint on speech and is unconstitutionally

vague. Nor do we decide their cognate unconstitutional

conditions theory, though for reasons explained infra at

note 36, we discuss the issues raised by that claim at

some length.

l4a

For these reasons, we will enter an Order declaring

Sections1712(a)(2) and 1721(b) of the Children’s Inter-

net Protection Act, codified at 20 U.S.C. § 9134(f) and

47 U.S.C. § 254(h)(6), respectively, to be facially invalid

under the First Amendment and permanently enjoining

the defendants from enforcing those provisions.

II. Findings of Fact

A. Statutory Framework

1. Nature and Operation of the E-rate and LSTA

Programs

In the Telecommunications Act of 1996 (“1996 Act”),

Congress directed the Federal Communications Com-

mission (“FCC”) to take the steps necessary to estab-

lish a system of support mechanisms to ensure the

delivery of affordable telecommunications service to all

Americans. This system, referred to as “universal

service,” is codified in section 254 of the Communi-

cations Act of 1934, as amended by the 1996 Act. See 47

U.S.C. § 254. Congress specified several groups as

beneficiaries of the universal service support mecha-

nism, including consumers in high-cost areas, low-in-

come consumers, schools and libraries, and rural health

care providers. See 47 U.S.C. § 254(h)(1). The exten-

sion of universal service to schools and libraries in

section 254(h) is commonly referred to as the Schools

and Libraries Program, or “E-rate” Program.

Under the E-rate Program, “({aJll telecommunications

carriers serving a geographic area shall, upon a bona

fide request for any of its services that are within the

definition of universal service . . ., provide such

services to elementary schools, secondary schools, and

libraries for educational purposes at rates less than the

amounts charged for similar services to other parties.”

15a

47 U.S.C. § 254(h)(1)(B). Under FCC regulations, pro-

viders of “interstate telecommunications” (with certain

exceptions, see 47 C. F. R. § 54.706(d)), must contribute a

portion of their revenue for disbursement among eligi-

ble carriers that are providing services to those groups

or areas specified by Congress in section 254. To be

eligible for the discounts, a library must: (1) be eligible

for assistance from a State library administrative

agency under the Library Services and Technology Act,

see infra; (2) be funded as an independent entity, com-

pletely separate from any schools; and (3) not be operat-

ing as a for-profit business. See 47 C. F. R. § 54.501(c).

Discounts on services for eligible libraries are set as a

percentage of the pre-discount price, and range from

20% to 90%, depending on a library’s level of economic

disadvantage and its location in an urban or rural area.

See 47 C.F.R. § 54.505. Currently, a library’s level of

economic disadvantage is based on the percentage of

students eligible for the national school lunch program

in the school district in which the library is located.

The Library Services and Technology Act (“LSTA”),

Subchapter II of the Museum and Library Services

Act, 20 U.S.C. § 9101 et seg., was enacted by Congress

in 1996 as part of the Omnibus Consolidated Appropria-

tions Act of 1997, Pub. L. No. 104-208. The LSTA es-

tablishes three grant programs to achieve the goal of

improving library services across the nation. Under

the Grants to States Program, LSTA grant funds are

awarded, inter alia, in order to assist libraries in

accessing information through electronic networks and

pay for the costs of acquiring or sharing computer

and telecommunications technologies. See 20

U.S.C. § 9141(a). Through the Grants to States pro-

gram, LSTA funds have been used to acquire and pay

16a

costs associated with Internet- accessible computers

located in libraries.

2. CIA

The Children's Internet Protection Act (“CIPA”) was

enacted as part of the Consolidated Appropriations Act

of 2001, which consolidated and enacted several ap-

propriations bills, including the Miscellaneous Appro-

priations Act, of which CIPA was a part. See Pub. L.

No. 106-554. CIPA addresses three distinct types of

federal funding programs: (1) aid to elementary and

secondary schools pursuant to Title III of the Elemen-

tary and Secondary Education Act of 1965, see CIPA

§ 1711 (amending Title 20 to add § 3601); (2) LSTA

grants to states for support of libraries, see CIPA

§ 1712 (amending the Museum and Library Services

Act, 20 U.S.C. § 9134); and (3) discounts under the E-

rate program, see CIPA § 1721(a) & (b) (both amending

the Communications Act of 1934, 47 U.S.C. § 254(h)).

Only sections 1712 and 1721(b) of CIPA, which apply to

libraries, are at issue in this case.

As explained in more detail below, CIPA requires

libraries that participate in the LSTA and E-rate

programs to certify that they are using software filters

on their computers to protect against visual depictions

that are obscene, child pornography, or in the case of

minors, harmful to minors. CIPA permits library

officials to disable the filters for patrons for bona fide

research or other lawful purposes, but disabling is not

permitted for minor patrons if the library receives

E-rate discounts.

a. CIPA’s Amendments to the E-rate Program

Section 1721(b) of CIPA imposes conditions on a

library’s participation in the E-rate program. A library

17a

“having one or more computers with Internet access

may not receive services at discount rates, CIPA

§ 1721(b) (codified at 47 U.S.C. § 254(h)(6)(A)(i)), unless

the library certifies that it is “enforcing a policy of

Internet safety that includes the operation of a tech-

nology protection measure with respect to any of its

computers with Internet access that protects against

access through such computers to visual depictions that

are (I) obscene; (II) child pornography; or (IIT) harm-

ful to minors,” and that it is — the operation —

techno tection measure during any use o

— — amen” CIPA § 1721(b) (codified at

47 U.S.C. § 254(h)\(6)(B)).2 CIPA defines a “technology

4 defines Im linor as any individual who has not at-

4 A CIPA § 1721(c) (codified at 47 US.C.

§ 254(h\(7(D)). CIPA further provides that “[oJbscene” has the

meaning given in 18 U.S.C. § 1460, and “child pornography” has

the meaning given in 18 U.S.C. § 2256. CIPA § 1721(c) (codified at

47 U.S.C. § SACHE) & FY. CIPA defines material that is

— — visual depiction

i i „graphie image file, or other i

ba-) taken as 4 whole and with respect to minors, appeals

to a prurient interest in nudity, sex, or excretion; (ii) depicts,

describes, or represents, in a patently offensive way with

respect to what is suitable for minors, an actual or simulated

sexual act or sexual contact, actual or simulated normal or

perverted sexual acts, or a lewd exhibition of the genitals; and

(iii) taken as a whole, lacks serious literary, artistic, political,

or scientific value as to minors.

18a

protection measure” as “a specific technology that

blocks or filters access to visual depictions that are

obscene, . . . child pornography, . . or harmful to

minors.” CIPA § 1703(b)(1) (codified at 47 U.S.C.

§ 254(h)(7)(1)).

To receive E-rate discounts, a library must also cer-

tify that filtering software is in operation during adult

use of the Internet. More specifically, with respect to

adults, a library must certify that it is “enforcing a

policy of Internet safety that includes the operation of a

technology protection measure with respect to any

of its computers with Internet access that protects

against access through such computers to visual depic-

tions that are—I) obscene; or (II) child pornography,”

and that it is “enforcing the operation of such technol-

ogy protection measure during any use of such com-

puters.” CIPA § 1721(b) (codified at 47 U.S.C.

§ 254(h)(6)(C)). Interpreting the statutory terms “any

use,” the FCC has concluded that “CIPA makes no

distinction between computers used only by staff and

those accessible to the public.” In re Federal State

Joint Board on Universal Service: Children Internet

Protection Act, CC Docket No. 96-45, Report and

Order, FCC 01-120, 1 30 (Apr. 5, 2001). ;

With respect to libraries receiving E-rate discounts,

CIPA further specifies that Lahn administrator, super-

visor, or other person authorized by the certifying

authority j may disable the technology protection

measure concerned, during use by an adult, to enable

19a

access for bona fide research or other lawful purpose.”

CIPA 172100 (codified at 47 U.S.C. § 25 40h) (6) D)).

b. CIPA’s Amendments to the LSTA Program

ion 1712 of CIPA amends the Museum and Lib-

vary 5 Servi — Act (20 U.S.C. § 9134(f)) to provide that

no funds made available under the Act “may be used to

purchase computers used to access the Internet, or to

pay for direct costs associated with accessing the Inter-

net,” unless such library “has in place” and is enforcing

“a policy of Internet safety that includes the operation

of a technology protection measure with respect to any

of its computers with Internet access that protects

against access through such computers to visual depic-

tions” that are “obscene” or “child pornography, and,

when the computers are in use by minors, also protects

against access to visual depictions that are Mr

to minors.” CIPA § 1712 (codified at 20 U.S.C.

§ 9134(f)(1)). Section 1712 contains definitions of “tech-

nology protection measure,” “obscene, child porno-

graphy,” and “harmful to minors, that are substan-

tially similar to those found in the provisions governing

the E-rate program. CIPA § 1712 (codified at 20 U.S.C.

§ 9134(f)(7)); see also supra note 2.

. inistrator

As under the E-rate program, “an adminis ,

supervisor or other authority may disable a technology

protection measure . . to enable access for bona fide

research or other lawful purposes.” CIPA § 1712

(codified at 20 U.S.C. § 9134(f)(3)). Whereas CIPA’s

amendments to the E-rate program permit disabling

for bona fide research or other lawful purposes only

during adult use, the LSTA provision permits disabling

for both adults and minors.

20a

B. Identity of the Plaintiffs

1. Library and Library Association Plaintiffs

Plaintiffs 4merican Library Association, Alaska Lib-

rary Association, California Library Association, Con-

necticut Library Association, Freedom to Read Foun-

dation, Maine Library Association, New England Li-

brary Association, New York Library Association, and

Wisconsin Library Association are non-profit organiza-

tions whose members include public libraries that

receive either E-rate discounts or LSTA funds for the

provision of Internet access. Because it is a pre-

requisite to_associational standing, we note that the

interests that these organizations seek to protect in this

litigation are central to their raison d’étre.

Plaintiffs Fort Vancouver Regional Library District,

in southwest Washington state; Multnomah County

Public Library, in Multnomah Count , Oregon; Norfolk

Public Library System, in N orfolk, Virginia; Santa Cruz

Public Library Joint Powers Authority, in Santa Cruz,

California; South Central Library System (“SCLS”),

centered in Madison, Wisconsin; and the Westchester

Library System, in Westchester County, New York,

are public library systems with branch offices in their

respective localities that provide Internet access to

their patrons.

The Fort Vancouver Regional Library District, for

over three years from 1999-2001, received $135,000 in

LSTA grants and $19,500 in E-rate discounts for Inter-

net access. The Multnomah County Public Library

received $70,000 in E-rate discounts for Internet access

this year, and has applied for $100,000 in E-rate

discounts for the upcoming year. The Norfolk Public

Library System received $90,000 in E-rate discounts

21a

ss this year, and has received a

—— LSTA erent to 1. computer labs in eight of

its libraries. The Santa Cruz Public Library Joint

Powers Authority received $20,560 in E-rate discounts

for Internet access in 2001-02. The SCLS received

between $3,000 and $5,000 this year in E-rate discounts

for Internet access.

The Fort Vancouver Regional Library District Board

is a public board whose members are appointed by

elected county commissioners. The Multnomah County

Library is a county department, whose board is ap-

pointed by the county chair and confirmed by-the 4

commissioners. The SCLS is an aggregation of -

independently governed statutory member pub —

libraries, whose relationship to SCLS is defined —

state law. The governing body of the SCLS is the

Library Board of Trustees, which consists of 20 mem-

bers nominated by county executives and ratified by

county boards of supervisors.

2. Patron and Patron Association Plaintiffs

; * 8 —

Plaintiffs Association of Community Organiza a

for Reform Now, Friends of the Philadelphia City Insti-

tute Library, and the Pennsylvania Alliance for

mocracy are nonprofit organizations whose members

techs — ee access the Internet at public

libraries that receive E-rate discounts or LSTA ——

for the provision of publie Internet access. We note —

the purpose of associational standing that the interes

that these organizations seek to protect in this litigation

are germane to their purposes. |

Plaintiffs Emmalyn Rood, Mark Brown, Elizabeth

Hrenda, C. Donald Weinberg, Sherron Dixon, by her

father and next friend Gordon Dixon, James Geringer,

22a

8 Tynesha Overby, by her next friend Carol

= 2 William J. Rosenbaum, Carolyn C.

—7 , and — Williams, by her mother and next

— haron er nard, are adults and minors who use

nternet at public libraries that, to the best of their

searching to be private. Although the library offered

patrons the option of using filteri ware

tering soft

did not use that option because she had had —

Plaintiff Mark Brown used

b N th

Philadelphia Free Library to 1 — 4

23a

clinical assistant professor of family medicine at Brown

University. Afraid to Ask. com's mission is to provide

detailed information on sensitive health issues, often of

a sexual nature, such as sexually transmitted diseases,

male and female genitalia, and birth control, sought by

people of all ages who would prefer to learn about sen-

sitive health issues anonymously, i.e., they are “afraid

to ask.” As part of its educational mission, Afraidto

Ask.com often uses graphic images of sexual anatomy

to convey information. Its primary audience is teens

and young adults. Based on survey data collected on

the site, half of the people visiting the site are under 24

years old and a quarter are under 18. AfraidtoAsk.com

is blocked by several leading blocking products as

containing sexually explicit content.

Plaintiff Alan Guttmacher Institute has a Web site

that contains information about its activities and objec-

tives, including its mission to protect the reproductive

choices of women and men. Plaintiff Planned Parent-

hood Federation of America, Inc. (“Planned Parent-

hood”) is a national voluntary organization in the field

of reproductive health care. Planned Parenthood owns

and operates several Web sites that provide a range of

information about reproductive health, from contracep-

tion to prevention of sexually transmitted diseases, to

finding an abortion provider, and to information about

the drug Mifepristone. Plaintiff Safersex.org is a Web

site that offers free educational information on how to

practice safer sex.

Plaintiff Ethan Interactive, Inc., d/b/a Out In

America, is an online content provider that owns and

operates 64 free Web sites for gay, lesbian, bisexual and

transgendered persons worldwide. Plaintiff PlanetOut

Corporation is an online content provider for gay,

24a

lesbian, bisexual and transgendered persons. Plaintiff

the Naturist Action Committee (“NAC”) is the non-

profit political arm of the Naturist Society, a private

organization that promotes a way of life characterized

by the practice of nudity. The NAC Web site provides

information about Naturist Society activities and about

state and local laws that may affect the rights of

Naturists or their ability to practice Naturism, and

includes nude photographs of its members.

Plaintiff Wayne L. Parker was the Libertarian

candidate in the 2000 U.S. Congressional election for

the Fifth District of Mississippi (and is running again in

2002). He publishes a Web site that communicates

information about his campaign and that provides infor-

mation about his political views and the Libertarian

Party to the public. Plaintiff Jeffrey Pollock was the

Republican candidate in the 2000 U.S. Congressional

election for the Third District of Oregon. He operates a

Web site that is now promoting his candidacy for

Congress in 2002“

8 The government challenges the standing of several of the

plaintiffs and the ripeness of their claims. These include all of the

Web site publishers and all of the individual library patrons.

Notwithstanding these objections, we are confident that the “case

or controversy” requirement of Article III, § 2 of the Constitution

is met by the existence of the plaintiff libraries that qualify for

LSTA and E-rate funding and the library associations whose

members qualify for such funding. These plaintiffs are faced with

the impending choice of either certifying compliance with CIPA by

July 1, 2002, or foregoing subsidies under the LSTA and E-rate

programs, and therefore clearly have standing to challenge the

constitutionality of the conditions to which they will be subject

should they accept the subsidies. We also note that the presence of

the Web site publishers and individual library patrons does not

affect our legal analysis or disposition of the case.

25a

C. The Internet

1. Background

As we noted at the outset, the Internet is a vast,

interactive medium consisting of a decentralized net-

work of computers around the world. The Internet

presents low entry barriers to anyone who wishes to

provide or distribute information. Unlike television,

cable, radio, newspapers, magazines or books, the

Internet provides an opportunity for those with access

to it to communicate with a worldwide audience at little

cost. At least 400 million people use the Internet

worldwide, and approximately 143 million Americans

were using the Internet as of September 2001. Nat’l

Telecomm.& Info. Admin., A Nation Online: How

Americans Are Expanding Their Use of the Internet

(February 2002), available at http:/;oww.ntia.doc.

gov/ntiahome/dn/.

The World Wide Web is a part of the Internet that

consists of a network of computers, called “Web

servers,” that host “pages” of content accessible via the

Hypertext Transfer Protocol or “HTTP.” Anyone with

a computer connected to the Internet can search for

and retrieve information stored on Web servers located

around the world. Computer users typically access the

Web by running a program called a “browser” on their

computers. The browser displays, as individual pages

on the computer screen, the various types of content

found on the Web and lets the user follow the con-

nections built into Web pages—called “hypertext

links,” “hyperlinks,” or “links”—to additional content.

Two popular browsers are Microsoft Internet Explorer

and Netscape Navigator.

26a

A “Web page” is one or more files a browser

graphically assembles to make a viewable whole when a

user requests content over the Internet. A Web page

may contain a variety of different elements, including

text, images, buttons, form fields that the user can fill

in, and links to other Web pages. A “Web site” is a

term that can be used in several different ways. It may

refer to all of the pages and resources available on a

particular Web server. It may also refer to all the

pages and resources associated with a particular or-

ganization, company or person, even if these are located

on different servers, or in a subdirectory on a single

server shared with other, unrelated sites. Typically, a

Web site has as an intended point of entry, a “home

page,” which includes links to other pages on the same

Web site or to pages on other sites. Online discussion

groups and chat rooms relating to a variety of subjects

are available through many Web sites.

Users may find content on the Web using engines

that search for requested keywords. In response to a

keyword request, a search engine will display a list of

Web sites that may contain relevant content and pro-

vide links to those sites. Search engines and directories

often return a limited number of sites in their search

results (e.g., the Google search engine will return only

2,000 sites in response to a search, even if it has found,

for example, 530,000 sites in its index that meet the

search criteria).

A user may also access content on the Web by typing

a URL (Uniform Resource Locator) into the address

line of the browser. A URL is an address that points to

some resource located on a Web server that is accessi-

ble over the Internet. This resource may be a Web site,

a Web page, an image, a sound or video file, or other

27a

resource. A URL can be either a numeric Internet

Protocol or “IP” address, or an alphanumeric “domain

name” address. Every Web server connected to the

Internet is assigned an IP address. A typical IP

address looks like “13.1.64.14.” Typing the URL “http://

13.1.64.14 /’ into a browser will bring the user to the

Web server that corresponds to that address. For con-

venience, most Web servers have alphanumeric domain

name addresses in addition to IP addresses. For

example, typing in “http:/;vww.paed.uscourts.gov” will

bring the user to the same Web server as typing in

“http://204.170.64.143.”

Every time a user attempts to access material

located on a Web server by entering a domain name

address into a Web browser, a request is made to a

Domain Name Server, which is a directory of domain

names and IP addresses, to “resolve,” or translate, the

domain name address into an IP address. That IP

address is then used to locate the Web server from

docile which content is being requested. A Web site

may be accessed by using either its domain name

address or its IP address.

A domain name address typically consists of several

parts. For example, the alphanumeric URL http://www.

paed.uscourts.gov/documents/opinions can be broken

down into three parts. The first part is the transfer

protocol the computer will use in accessing the content

(e.g., “http” for Hypertext Transfer Protocol); next is

the name of the host server on which the information is

stored (e.g., www.paed.uscourts.gov); and then the

name of the particular file or directory on that server

(e. g., documents / opinions).

A ean Wile quae — with mave then

one URL. For example, the URLs http://www.new

28a

yorktimes.com and http://www.nytimes.com will both

take the user to the New York Times home page. The

topmost directory in a Web site is often referred to as

that Web site’s root directory or root URL. For

example, in http:/;www.paed.uscourts.gov/documents,

the root URL is http:// www.paed.uscourts.gov. There

may be hundreds or thousands of pages under a single

root URL, or there may be one or only a few.

There are a number of Web hosting companies that

maintain Web sites for other businesses and indivi-

duals, which can lead to vast amounts of diverse content

being located at the same IP address. Hosting services

are offered either for a fee, or in some cases, for free,

allowing any individual with Internet access to create a

Web site. Some hosting services are provided through

the process of “IP-based hosting,” where each domain

name is assigned a unique IP number. For example,

www.baseball.com might map to the IP address

“10.3.5.9” and www.XXX.com might map to the IP

address “10.0.42.5.” Other hosting services are pro-

vided through the process of “name-based hosting,”

where multiple domain name addresses are mapped to a

single IP address. If the hosting company were using

this method, both www.baseball.com and www.XXX.

com could map to a single IP address, e.g., “10.3.5.9.”

As a result of the “name-based hosting” process, up to

tens of thousands of pages with heterogeneous content

may share a single IP address.

2. The Indexable Web, the “Deep Web”; Their size

and Rates of Growth and Change

The universe of content on the Web that could be

indexed, in theory, by standard search engines is

known as the “publicly indexable Web.” The publicly

indexable Web is limited to those pages that are

29a

accessible by following a link from another Web page

that is recognized by a search engine. This limitation

exists because online indexing techniques used by

popular search engines and directories such as Yahoo,

Lycos and AltaVista, are based on “spidering” tech-

nology, which finds sites to index by following links

from site to site in a continuous search for new content.

If a Web page or site is not linked by others, then

spidering will not discover that page or site.

Furthermore, many larger Web sites contain instruc-

tions, through software, that prevent spiders from

investigating that site, and therefore the contents of

such sites also cannot be indexed using spidering

technology. Because of the vast size and decentralized

structure of the Web, no search engine or directory

indexes all of the content on the publicly indexable

Web. We credit current estimates that no more than

50% of the content currently on the publicly indexable

Web has been indexed by all search engines and

directories combined. No currently available method or

combination of methods for collecting URLs can collect

the addresses of all URLs on the Web.

The portion of the Web that is not theoretically

indexable through the use of “spidering” technology,

because other Web pages do not link to it, is called the

“Deep Web.” Such sites or pages can still be made

publicly accessible without being made publicly index-

able by, for example, using individual or mass emailings

(also known as “spam”) to distribute the URL to

potential readers or customers, or by using types of

Web links that cannot be found by spiders but can be

seen and used by readers. “Spamming” is a common

method of distributing to potential customers links to

sexually explicit content that is not indexable.

30a

Because the Web is decentralized, it is impossible to

say exactly how large it is. A 2000 study estimated a

total of 7.1 million unique Web sites, which at the Web’s

historical rate of growth, would have increased to 11

million unique sites as of September 2001. Estimates of

the total number of Web pages vary, but a figure of 2

billion is a reasonable estimate of the number of Web

pages that can be reached, in theory, by standard

search engines. We need not make a specific finding as

to a figure, for by any measure the Web is extremely

vast, and it is constantly growing. The indexable Web

is growing at a rate of approximately 1.5 million pages

per day. The size of the un-indexable Web, or the

“Deep Web,” while impossible to determine precisely, is

estimated to be two to ten times that of the publicly

indexable Web.

In addition to growing rapidly, Web pages and sites

are constantly being removed, or changing their con-

tent. Web sites or pages can change content without

changing their domain name addresses or IP addresses.

Individual Web pages have an average life span of

approximately 90 days.

3. The Amount of Sexually Explicit Material on the

Web

There is a vast amount of sexually explicit material

available via the Internet and the Web. Sexually ex-

plicit material on the Internet is easy to access using

any public search engine, such as, for example, Google

or AltaVista. Although much of the sexually explicit

material available on the Web is posted on commercial

sites that require viewers to pay in order to gain access

to the site, a large number of sexually explicit sites may

be accessed for free and without providing any registra-

31a

tion information. Most importantly, some Web sites

that contain sexually explicit content have innocuous

domain names and therefore can be reached acciden-

tally. A commonly cited example is http://www. white-

house.com. Other innocent-sounding URLs that re-

trieve graphic, sexually explicit depictions include

http://www.boys.com, http:/hwww.girls.com, http:// www.

coffeebeansupply.com, and http://www.BookstoreUSA.

com. Moreover, commercial Web sites that contain

sexually explicit material often use a technique of

attaching pop-up windows to their sites, which open

new windows advertising other sexually explicit sites

without any prompting by the user. This technique

makes it difficult for a user quickly to exit all of the

pages containing sexually explicit material, whether he

or she initially accessed such material intentionally or

not.

The percentage of Web pages on the indexed Web

containing sexually explicit content is relatively small.

Recent estimates indicate that no more than 1-2% of

the content on the Web is pornographic or sexually

explicit. However, the absolute number of Web sites

offering free sexually explicit material is extremely

large, approximately 100,000 sites.

D. American Public Libraries

The more than 9,000 public libraries in the United

States are typically funded (at least in large part) by

state or local governments. They are frequently

overseen by a board of directors that is either elected

or is appointed by an elected official or a body of elected

officials. We heard testimony from librarians and

library board members working in eight public library

systems in different communities across the country,

some of whom are also plaintiffs in this case. They

32a

hailed from the following library systems: Fort Van-

couver, Washington; Fulton County, Indiana; Green-

ville, South Carolina; a regional consortium of libraries

centered in Madison, Wisconsin; Multnomah County,

Oregon; Norfolk, Virginia; Tacoma, Washington; and

Westerville, Ohio. The parties also took depositions

from several other librarians and library board mem-

bers —— 1 during the trial, and submitted

a number of other documents re ing indivi

2 i garding individual

1. The Mission of Public Libraries, and Their

Reference and Collection Development Practices

American public libraries operate in a wide variety of

communities, and it is not surprising that they do not all

view their mission identically. Nor are their practices

uniform. Nevertheless, they generally share a common

mission—to provide patrons with a wide range of

information and ideas. a

Public libraries across the country have endorsed the

American Library Association’s (“ALA”) “Library Bill

of Rights” and/or “Freedom to Read Statement,” in-

cluding every library testifying on behalf of the defen-

dants in this case. The “Library Bill of Rights,” first

adopted by the ALA in 1948, provides, among other

things, that books and other library resources should

be provided for the interest, information, and enlighten-

ment of all people of the community the library serves.”

It also states that libraries “should provide materials

and information presenting all points of view on current

and historical issues” and that library materials “should

not be proscribed or removed because of partisan or

doctrinal disapproval.”

33a

The ALA’s “Freedom to Read” statement, adopted in

1953 and most recently updated in July 2000, states,

among other things, that “[iJt is in the public interest

for publishers and librarians to make available the

widest diversity of views and expressions, including

those that are unorthodox or unpopular with the

majority.” It also states that ſiſt is the responsibility

of ... librarians . . to contest encroachments upon

thle] freedom [to read] by individuals or groups seeking

to impose their own standards or tastes upon the

community at large.”

Public libraries provide information not only for

educational purposes, but also for recreational, profes-

sional, and other purposes. For example, Ginnie

Cooper, Directo of the Multnomah County Library,

testified that sorac of the library’s most popular items

include video tapes of the British Broadcasting Cor-

poration’s “Fawlty Towers” series, and also print and

“books on tape” versions of science fiction, romance,

and mystery novels. Many public libraries include

sexually explicit materials in their print collection, such

as The Joy of Sex and The Joy of Gay Sex. Very few

public libraries, however, collect more graphic sexually

explicit materials, such as XXX-rated videos, or

Hustler magazine. ‘

The mission of public librarians is to provide their

patrons with a wide array of information, and they

surely do so. Reference librarians across America

4 The OCLC database, a cooperative cataloging service estab-

lished to facilitate interlibrary loan requests, includes 40 million

catalog records from approximately 48,000 libraries of all types

worldwide. Slightly more than 400 of the libraries in the OCLC

database are listed as carrying Playboy in their collections, while

only eight subscribe to Hustler.

34a

answer more than 7 million questions weekly. If a

patron has a specialized need for information not

available in the public library, the professional librarian

will use a reference-interview to find out what infor-

mation is needed to help the user, including the purpose

for which an item willbe used. Reference librarians are

trained to assist patrons without judging the patron’s

purpose in seeking information, or the content of the

information that the patron is seeking.

Many public libraries routinely provide patrons with

access to materials not in their collections through the

use of bibliographic access tools and interlibrary loan

programs. Public libraries typically will assist patrons

in obtaining access to all materials except those that are

illegal, even if they do not collect those materials in

their physical collection. In order to provide this

access, a librarian may attempt to find material not

included in the library’s own collection in other libraries

in the system, through interlibrary loan, or through a

referral, perhaps to a government agency or a com-

mercial bookstore. Interlibrary loan is expensive, how-

ever, and is therefore used infrequently.

Public librarians also apply professional standards to

their collection development practices. Public libraries

generally make material selection decisions and frame

policies governing collection development at the local

level. Collection development is a key subject in the

curricula of Masters of Library Science programs and is

defined by certain practices. In general, professional

standards guide public librarians to build, develop and

peo oa that have certain characteristics, such

as ce in its coverage and requisite and appropriate

quality. To this end, the goal of library collections is not

universal coverage, but rather to find those materials

35a

that would be of the greatest lireet benefit or interest

to the community. In making selection decisions,

librarians consider criteria including the content of the

material, its accuracy, the title’s niche in relation to the

rest of the collection, the authority of the author, the

publisher, the work’s presentation, and how it compares

with other material available in the same genre or on

the same subject.

In pursuing the goal of achieving a balanced collec-

tion that serves the needs and interests of their pa-

trons, librarians generally have a fair amount of auton-

omy, but may also be guided by a library’s collection

development policy. These collection development

policies are often drawn up in conjunction with the

libraries’ governing boards and with representatives

from the community, and may be the result of public

hearings, discussions and other input.

Although many librarians use selection aids, such as

review journals and bibliographies, as a guide to the

quality of potential acquisitions, they do not generally

delegate their selection decisions to parties outside of

the public library or its governing body. One limited

exception is the use of third-party vendors or approval

plans to acquire print and video resources. In such

ments, third-party vendors provide materials

based on the library’s description of its collection devel-

opment criteria. The vendor sends materials to the

library, and the library retains the materials that meet

its collection development needs and returns the

materials that do not. Even in this arrangement,

however, the librarians still retain ultimate control over

their collection development and review all of the

materials that enter their library’s collection.

36a

2. The Internet in Public Libraries

The vast majority of public libraries offer Internet

access to their patrons. According to a recent report by

the US. National Commission on Libraries and Infor-

mation Science, approximately 95% of all public

libraries provide public access to the Internet. John C.

Bertot & Charles R. McClure, Public Libraries and the

Internet 2000: Summary Findings and Data Tables,

Report to National Commission on Libraries and Infor-

mation Science, at 3. The Internet vastly expands the

amount of information available to patrons of public

libraries. The widespread availability of Internet

access in public libraries is due, in part, to the availabil-

ity of public funding, including state and local funding

and the federal funding programs regulated by CIPA.

Many libraries face a large amount of patron demand

for their Internet services. At some libraries, patron

demand for Internet access during a given day exceeds

the supply of computer terminals with access to the

Internet. These libraries use sign-in and time limit pro-

cedures and/or establish rules regarding the allowable

uses of the terminals, in an effort to ration their

computer resources. For example, some of the libraries

whose librarians testified at trial prohibit the use of

email and chat functions on their public Internet

terminals.

Public libraries play an important role in providing

Internet access to citizens who would not otherwise

possess it. Of the 143 million Americans using the

Internet, approximately 10%, or 14.3 million people,

access the Internet at a public library. Internet access

at public libraries is more often used by those with

lower incomes than those with higher incomes. About

20.3% of Internet users with household family income of

37a

less than $15,000 per year use public libraries for Inter-

net access. Approximately 70% of libraries serving

communities with poverty levels in excess of 40%

receive E-rate discounts.

a. Internet Use Policies in Public Libraries

Approximately 95% of libraries with public Internet

access have some form of “acceptable use” policy or

“Internet use” policy governing patrons’ use of the

Internet. These policies set forth the conditions under

which patrons are permitted to access and use the

library’s Internet resources. These policies vary

widely. Some of the less restrictive policies, like those

held by Multnomah County Library and Fort Van-

couver Regional Library, do not prohibit adult patrons

from viewing sexually explicit materials on the Web, as

long as they do so at terminals with privacy screens or

recessed monitors, which are designed to prevent other

patrons from seeing the material that they are viewing,

and as long as it does not violate state or federal law to

do so. Other libraries prohibit their patrons from

viewing all “sexually explicit” or “sexually graphic”

materials.

Some libraries prohibit the viewing of materials that

are not necessarily sexual, such as Web pages that are

“harmful to minors,” “offensive to the public,” “objec-

tionable,” “racially offensive,” or simply “inappropri-

ate.” Other libraries restrict access to Web sites that

the library just does not want to provide, even though

the sites are not necessarily offensive. For example,

the Fulton County Public Library restricts access to

the Web sites of dating services. Similarly, the Tacoma

Public Library’s policy does not allow patrons to use

38a

the library’s Internet terminals for personal email, for

online chat, or for playing games.

In some cases, libraries instituted Internet use

policies after having experienced specific problems,

whereas in other cases, libraries developed detailed

Internet use policies and regulatory measures (such as

using filtering software) before ever offering public

Internet access. Essentially four interests motivate

libraries to institute Internet use policies and to apply

the methods described above to regulate their patrons’

use of the Internet.

First, libraries have sought to protect patrons (espe-

cially children) and staff members from accidentally

viewing sexually explicit images, or other Web pages

containing content deemed harmful, that other patrons

are viewing on the Internet. For example, some

librarians who testified described situations in which

patrons left sexually explicit images minimized on an

Internet terminal so that the next patron would see

them when they began using it, or in which patrons

printed sexually explicit images from a Web site and

left them at a public printer.

Second, libraries have attempted to protect patrons

from unwittingly or accidentally accessing Web pages

that they do not wish to see while they are using the

Internet. For example, the Memphis Shelby County

(Tennessee) Public Library’s Internet use policy states

that the library “employs filtering technology to reduce

the possibility that customers may encounter objec-

tionable content in the form of depictions of full nudity

and sexual acts.”

Third, libraries have sought to keep patrons (again,

especially children) from intentionally accessing sexu-

39a

ally explicit materials or other materials that the

library deems inappropriate. For example, a study of

the Tacoma Public Library’s Internet use logs for the

year 2000 showed that users between the ages of 11 and

15 accounted for 41% of the filter blocks that occurred

on library computers. The study, which we credit,

concluded that children and young teens were actively

seeking to access sexually explicit images in the library.

The Greenville Library’s Board of Directors was par-

ticularly concerned that patrons were accessing ob-

scene materials in the public library in violation of

South Carolina’s obscenity statute.

Finally, some libraries have regulated patrons’ Inter-

net use to attempt to control patrons’ inappropriate (or

illegal) behavior that is thought to stem from viewing

Web pages that contain sexually explicit materials or

content that is otherwise deemed unacceptable.

We recognize the concerns that led several of the

public libraries whose librarians and board members

testified in this case to start using Internet filtering

software. The testimony of the Chairman of the Board

of the Greenville Public Library is illustrative. In

December 1999, there was considerable local press cov-

erage in Greenville concerning adult patrons who rou-

tinely used the library to surf the Web for pornography.

In response to public outcry stemming from the news-

paper report, the Board of Trustees held a special board

meeting to obtain information and to communicate with

the public concerning the library’s provision of Internet

access. At this meeting, the Board learned for the first

time of complaints about children being exposed to

pornography that was displayed on the library’s

Internet terminals.

40a

In late January to early February of 2000, the library

installed privacy screens and recessed terminals in an

effort to restrict the display of sexually explicit Web

sites at the library. In February, 2000, the Board

informed the library staff that they were expected to be

familiar with the South Carolina obscenity statute and

to enforce the policy prohibition on access to obscene

materials, child pornography, or other materials pro-

hibited under applicable local, state, and federal laws.

Staff were told that they were to enforce the policy by

means of a “tap on the shoulder.” Prior to adopting its

current Internet Use Policy, the Board adopted an

“Addendum to Current Internet Use Policy.” Under

the policy, the Board temporarily instituted a two-hour

time limit per day for Internet use; reduced sub-

stantially the number of computers with Internet

access in the library; reconfigured the location of the

computers so that librarians had visual contact with all

Internet- accessible terminals; and removed the privacy

screens from terminals with Internet access.

Even after the Board implemented the privacy

screens and later the “tap-on-the-shoulder” policy com-

bined with placing terminals in view of librarians, the

library experienced a high turnover rate among refer-

ence librarians who worked in view of Internet

terminals. Finding that the policies that it had tried did

not prevent the viewing of sexually explicit materials in

the library, the Board at one point considered dis-

continuing Internet access in the library. The Board

finally concluded that the methods that it had used to

regulate Internet use were not sufficient to stem the

behavioral problems that it thought were linked to the

availability of pornographic materials in the library. As

a result, it implemented a mandatory filtering policy.

41a

We note, however, that none of the libraries prof-

fered by the defendants presented any systematic re-

cords or quantitative comparison of the amount of

criminal or otherwise inappropriate behavior that oc-

curred in their libraries before they began using Inter-

net filtering software compared to the amount that

happened after they installed the software. The plain-

tiffs’ witnesses also testified that because public librar-

ies are public places, incidents involving inappropriate

behavior in libraries (sexual and otherwise) existed

long before libraries provided access to the Internet.

b. Methods for Regulating Internet Use

The methods that public libraries use to regulate

Internet use vary greatly. They can be organized into

four categories: (1) channeling patrons’ Internet use;

(2) separating patrons so that they will not see what

other patrons are viewing; (3) placing Internet ter-

minals in public view and having librarians observe

patrons to make sure that they are complying with the

library’s Internet use policy; and (4) using Internet

filtering software.

The first category—channeling patrons’ Internet

use—frequently includes offering training to patrons on

how to use the Internet, including how to access the

information that they want and to avoid the materials

that they do not want. Another technique that some

public libraries use to direct their patrons to pages that

the libraries have determined to be accurate and

valuable is to establish links to “recommended Web

sites” from the public library’s home page (i.e., the page

that appears when patrons begin a session at one of the

library’s public Internet terminals). Librarians select

these recommended Web sites by using criteria similar

to those employed in traditional collection development.

42a

However, unless the library determines otherwise,

selection of these specific sites does not preclude

patrons from attempting to access other Internet Web

sites.

Libraries may extend the “recommended Web sites”

method further by limiting patrons’ access to only those

Web sites that are reviewed and selected by the

library’s staff. For example, in 1996, the Westerville,

Ohio Library offered Internet access to children

through a service called the “Library Channel.” This

service was intended to be a means by which the library

could organize the Internet in some fashion for pre-

sentation to patrons. Through the Library Channel,

the computers in the children’s section of the library

were restricted to 2,000 to 3,000 sites selected by

librarians. After three years, Westerville stopped

using the Library Channel system because it overly

constrained the children’s ability to access materials on

the Internet, and because the library experienced

several technical problems with the system.

Public libraries also use several different techniques

to separate patrons during Internet sessions so that

they will not see what other patrons are viewing. The

simplest way to achieve this result is to position the

_ library’s public Internet terminals so that they are

located away from traffic patterns in the library (and

from other terminals), for example, by placing them so

that they face a wall. This method is obviously con-

strained by libraries’ space limitations and physical

layout. Some libraries have also installed privacy

screens on their public Internet terminals. These

screens make a monitor appear blank unless the viewer

43a

is looking at it head-on.’ Although the Multnomah and

Fort Vancouver Libraries submitted records showing

that they have received few complaints regarding

patrons’ unwilling exposure to materials on the Inter-

net, privacy screens do not always prevent library

patrons or employees from inadvertently seeing the

materials that another patron is viewing when passing

directly behind a terminal. They also have the draw-

back of making it difficult for patrons to work together

at a single terminal, or for librarians to assist patrons at

terminals, because it is difficult for two people to stand

side by side and view a screen at the same time. Some

library patrons also find privacy screens to be a

hindrance and have attempted to remove them in order

to improve the brightness of the screen or to make the

view better. N

Another method that libraries use to prevent patrons

from seeing what other patrons are viewing on their

terminals is the installation of recessed monitors.“

Recessed monitors are computer screens that sit below

the level of a desk top and are viewed from above.

Although recessed monitors, especially when combined

with privacy screens, eliminate almost all of the pos-

sibility of a patron aceidentally viewing the contents on

another patron’s screen, they suffer from the same

drawbacks as privacy screens, that is, they make it

5 Fort Vancouver Regional Library, for example, combines the

methods of strategically placing terminals in low traffic areas and

using privacy screens. A section headed “Confidentiality and Pri-

vacy” on the library's home page states: “in order to protect the

privacy of the user and the interests of other library patrons, the

library will attempt to minimize unintentional viewing of the Inter-

net. This will be done by use of privacy screens, and by judicious

placement of the terminals and other appropriate means.”

44a

difficult for patrons to work together or with a librarian

at a single terminal. Some librarians also testified that

recessed monitors are costly, but did not indicate how

expensive they are compared to privacy screens or

filtering software. A related technique that some pub-

lic libraries use is to create a separate children’s Inter-

net viewing area, where no adults except those accom-

panying children in their care may use the Internet

terminals. This serves the objective of keeping children

from inadvertently viewing materials appropriate only

for adults that adults may be viewing on nearby

terminals.

A third set of techniques that public libraries have

used to enforce their Internet use policies takes the

opposite tack from the privacy screens/recessed moni-

tors approach by placing all of the library’s public

Internet terminals in prominent and visible locations,

such as near the library’s reference desk. This

approach allows librarians to enforce their library’s

Internet use policy by observing what patrons are

viewing and employing the tap-on-the-shoulder policy.

Under this approach, when patrons are viewing materi-

als that are inconsistent with the library’s policies, a

library staff member approaches them and asks them to

view something else, or may ask them to end their

Internet session. A patron who does not comply with

these requests, or who repeatedly views materials not

permitted under the library’s Internet use policy, may

have his or her Internet or library privileges suspended

or revoked. But many librarians are uncomfortable with

approaching patrons who are viewing sexually explicit

images, finding confrontation unpleasant. Hence some

libraries are reluctant to apply the tap-on-the-shoulder

policy.

45a

The fourth category of methods that public libraries

employ to enforce their Internet use policies, and the

one that gives rise to this case, is the use of Internet

filtering software. According to the June 2000 Survey

of Internet Access Management in Public Libraries,

approximately 7% of libraries with public Internet

access had mandated the use of blocking programs by

adult patrons. Some public libraries provide patrons

with the option of using a blocking program, allowing

patrons to decide whether to engage the program when

they or their children access the Internet. Other public

libraries require their child patrons to use filtering

software; but not their adult patrons.

Filtering software vendors sell their products on a

subscription basis. The cost of a subscription varies

with the number of computers on which the filtering

software will be used. In 2001, the cost of the Cyber

Patrol filtering software was $1,950 for 100 terminal

licenses. The Greenville County Library System pays

$2,500 per year for the N2H2 filtering software, and a

subscription to the Websense filter costs Westerville

Public Library approximately $1,200 per year.

No evidence was presented on the cost of privacy

screens, recessed monitors, and the tap-on-the-

shoulder policy, relative to the costs of filtering soft-

ware. Nor did any of the libraries proffered by the

government present any quantitative evidence on the

relative effectiveness of use of privacy screens to pre-

vent patrons from being unwillingly exposed to sexu-

ally explicit material, and the use of filters, discussed

below. No evidence was presented, for example, com-

paring the number of patron complaints in those

libraries that have tried both methods.

46a

The librarians who testified at trial whose libraries

use Internet filtering software all provide methods by

which their patrons may ask the library to unblock

specific Web sites or pages. Of these, only the Tacoma

Public Library allows patrons to request that a URL be

unblocked without providing any identifying informa-

tion; Tacoma allows patrons to request a URL by

sending an email from the Internet terminal that the

patron is using that does not contain a return email

address for the user. David Biek, the head librarian at

the Tacoma Library’s main branch, testified at trial

that the library keeps records that would enable it to

know which patrons made unblocking requests, but

does not use that information to connect users with

their requests. Biek also testified that he periodically

scans the library’s Internet use logs to search for:

(1) URLs that were erroneously blocked, so that he

may unblock them; or (2) URLs that should have been

blocked, but were not, in order to add them to a blocked

category list. In the course of scanning the use logs,

Biek has also found what looked like attempts to access

child pornography. In two cases, he communicated his

findings to law enforcement and turned over the logs in

response to a subpoena.

At all events, it takes time for librarians to make

decisions about whether to honor patrons’ requests to

unblock Web pages. In the libraries proffered by the

defendants, unblocking decisions sometimes take

between 24 hours and a week. Moreover, none of these

libraries allows unrestricted access to the Internet

pending a determination of the validity of a Web site

blocked by the blocking programs. A few of the defen-

dants’ proffered libraries represented that individual

librarians would have the discretion to allow a patron to

47a

have full Internet access on a staff computer upon

request, but none claimed that allowing such access was

mandatory, and patron access is supervised in every

instance. None of these libraries makes differential

unblocking decisions based on the patrons’ age. Un-

blocking decisions are usually made identically for

adults and minors. Unblocking decisions even for

adults are usually based on suitability of the Web site

for minors.

It is apparent that many patrons are reluctant or

unwilling to ask librarians to unblock Web pages or

sites that contain only materials that might be deemed

personal or embarrassing, even if they are not sexually

explicit or pornographic. We credit the testimony of

Emmalyn Rood, discussed above, that she would have

been unwilling as a young teen to ask a librarian to dis-

able filtering software so that she could view materials

concerning gay and lesbian issues. We also credit the

testimony of Mark Brown, who stated that he would

have been too embarrassed to ask a librarian to disable

filtering software if it had impeded his ability to

research treatments and cosmetic surgery options for

his mother when she was diagnosed with breast cancer.

The pattern of patron requests to unblock specific

URLs in the various libraries involved in this case also

confirms our finding that patrons are largely unwilling

to make unblocking requests unless they are permitted

to do so anonymously. For example, the Fulton County

Library receives only about 6 unblocking requests each

year, the Greenville Public Library has received only 28

unblocking requests since August 21, 2000, and the

Westerville, Ohio Library has received fewer than 10

unblocking requests since 1999. In light of the fact that

a substantial amount of overblocking occurs in the-

48a

severy libraries, see infra Subsection II.E.4, we find

that the lack of unblocking requests in these libraries

does not reflect the effectiveness of the filters, but

rather reflects patrons’ reluctance to ask librarians to

unblock sites.

E. Internet Filtering Technology

1. What Is Filtering Software, Who Makes It, and

What Does It Do?

Commercially available products that can be config-

ured to block or filter access to certain material on the

Internet are among the “technology protection mea-

sures” that may be used to attempt to comply with

CIPA. There are numerous filtering software products

available commercially. Three network-based filtering

products—SurfControl’s Cyber Patrol, N2H2’s Bess/

i2100, and Secure Computing’s SmartFilter—currently

have the lion’s share of the public library market. The

parties in this case deposed representatives from these

three companies. Websense, another network-based

blocking product, is also currently used in the public

library market, and was discussed at trial.

Filtering software may be installed either on an

individual computer or on a computer network. Net-

work-based filtering software products are designed for

use on a network of computers and funnel requests for

Internet content through a centralized network device.

Of the various commercially available blocking pro-

ducts, network-based products are the ones generally

marketed to institutions, such as public libraries, that

provide Internet access through multiple terminals.

Filtering programs function in a fairly simple way.

When an Internet user requests access to a certain

Web site or page, either by entering a domain name or

49a

IP address into a Web browser, or by clicking on a link,

the filtering software checks that domain name or IP

address against a previously compiled “control list” that

may contain up to hundreds of thousands of URLs. The

three companies deposed in this case have control lists

containing between 200,000 and 600,000 URLs. These

lists determine which URLs will be blocked.

Filtering software companies divide their control

lists into multiple categories for which they have

created unique definitions. SurfControl uses 40 such

categories, N2H2 uses 35 categories (and seven “excep-

tion” categories), Websense uses 30 categories, and

Secure Computing uses 30 categories. Filtering soft-

ware customers choose which categories of URLs they

wish to enable. A user “enables” a category in a filter-

ing program by configuring the program to block all of

the Web pages listed in that category.

The following is a list of the categories offered by

each of these four filtering programs. SurfControl’s

Cyber Patrol offers the following categories: Adult/

Sexually Explicit; Advertisements; Arts & Entertain-

ment; Chat; Computing & Internet; Criminal Skills;

Drugs, Alcohol & Tobacco; Education; Finance & In-

vestment; Food & Drink; Gambling; Games; Glamour &

Intimate Apparel; Government & Politics; Hacking;

Hate Speech; Health & Medicine; Hobbies & Recrea-

tion; Hosting Sites; Job Search & Career Development;

Kids’ Sites; Lifestyle & Culture; Motor Vehicles; News;

Personals & Dating; Photo Searches; Real Estate;

Reference; Religion; Remote Proxies; Sex Education;

Search Engines; Shopping; Sports; Streaming Media;

Travel; Usenet News; Violence; Weapons; and Web-

based Email.

50a

N2H2 offers the following categories: Adults Only;

Alcohol; Auction; Chat; Drugs; Electronic Commerce;

Employment Search; Free Mail; Free Pages; Gambling;

Games; Hate/Discrimination; Illegal; Jokes; Lingerie;

Message/Bulletin Boards; Murder/Suicide; News; Nu-

dity; Personal Information; Personals; Pornography;

Profanity; Recreation/Entertainment; School Cheating

Information; Search Engines; Search Terms; Sex;

Sports; Stocks; Swimsuits; Tasteless/Gross; Tobacco;

Violence; and Weapons. The “Nudity” category pur-

ports to block only “non-pornographic” images. The

“Sex” category is intended to block only those depic-

tions of sexual activity that are not intended to arouse.

The “Tasteless/Gross” category includes contents such

as “tasteless humor” and “graphic medical or accident

scene photos.” Additionally, N2H2 offers seven “excep-

tion categories.” These exception categories include

Education, Filtered Search Engine, For Kids, History,

Medical, Moderated, and Text/Spoken Only. When an

exception category is enabled, access to any Web site or

page via a URL associated with both a category and an

exception, for example, both “Sex” and “Education,”

will be allowed, even if the customer has enabled the

product to otherwise block the category “Sex.” As of

November 15, 2001, of those Web sites categorized by

N2H2 as “Sex,” 3.6% were also categorized as “Educa-

tion,” 2.9% as “Medical,” and 1.6% as “History.”

Websense offers the following categories: Abortion

Advocacy; Advocacy Groups; Adult Material; Business

& Economy; Drugs; Education; Entertainment; Gamb-

ling; Games; Government; Health; Illegal/Questionable;

Information Technology; Internet Communication; Job

Search; Militancy/Extremist; News & Media; Produc-

tivity Management; Bandwidth Management; Racism/

5la

Hate; Religion; Shopping; Society & Lifestyle; Special

Events; Sports; Tasteless; Travel; Vehicles; Violence;

and Weapons. The “Adult” category includes “full or

partial nudity of individuals,” as well as sites offering

“light adult humor and literature” and As lexually ex-

plicit language.” The “Sexuality/Pornography” cate-

gory includes, inter alia, “hard-core adult humor and

literature” and As lexually explicit language.” The

“Tasteless” category includes hard- to-stomach sites,

ineluding offensive, worthless or useless sites, gro-

tesque or lurid depictions of bodily harm.“ The Hack-

ing” category blocks “sites providing information on or

promoting illegal or questionable access to or use of

communications equipment and/or software.”

SmartFilter offers the following categories:

Anonymizers/Translators; Art & Culture; Chat; Crimi-

nal Skills; Cults/Occult; Dating; Drugs; Entertainment;

Extreme/Obscene/Violence; Gambling; Games; General

News; Hate Speech; Humor; Investing; Job Search;

Lifestyle; Mature; MP3 Sites; Nudity; On-line Sales;

Personal Pages; Politics, Opinion & Religion; Portal

Sites; Self Help/Health; Sex; Sports; Travel; Usenet

News; and Webmail.

Most importantly, no category definition used by

filtering software companies is identical to CIPA’s

definitions of visual depictions that are obscene, child

pornography, or harmful to minors. And category

definitions and categorization decisions are made with-

out reference to local community standards. Moreover,

there is no judicial involvement in the creation of

filtering software companies’ category definitions and

no judicial determination is made before these com-

panies categorize a Web page or site.

52a

Each filtering software company associates each

URL in its control list with a “tag” or other identifier

that indicates the company’s evaluation of whether the

content or features of the Web site or page accessed via

that URL meets one or more of its category definitions.

If a user attempts to access a Web site or page that is

blocked by the filter, the user is immediately presented

with a screen that indicates that a block has occurred as

a result of the operation of the filtering software.

These “denial screens” appear only at the point that a

user attempts to access a site or page in an enabled

category.

All four of the filtering programs on which evidence

was presented allow users to customize the category

lists that exist on their own PCs or servers by adding

or removing specific URLs. For example, if a public

librarian charged with administering a library’s Inter-

net terminals comes across a Web site that he or she

finds objectionable that is not blocked by the filtering

program that his or her library is using, then the

librarian may add that URL to a category list that

exists only on the library’s network, and it would

thereafter be blocked under that category. Similarly, a

customer may remove individual URLs from category

lists. Importantly, however, no one but the filtering

companies has access to the complete list of URLs in

any category. The actual URLs or IP addresses of the

Web sites or pages contained in filtering software

vendors’ category lists are considered to be proprietary

information, and are unavailable for review by cus-

53a

tomers or the general public, including the proprietors

of Web sites that are blocked by filtering software.

Filtering software companies do not generally notify

the proprietors of Web sites when they block their

sites. The only way to discover which URLs are

blocked and which are not blocked by any particular

filtering company is by testing individual URLs with

filtering software, or by entering URLs one by one into

the “URL checker” that most filtering software com-

panies provide on their Web sites. Filtering software

companies will entertain requests for recategorization

from proprietors of Web sites that discover their sites

are blocked. Because new pages are constantly being

added to the Web, filtering companies provide their

customers with periodic updates of category lists. Once

a particular Web page or site is categorized, however,

filtering companies generally do not re-review the con-

tents of that page or site unless they receive a request

to do so, even though the content on individual Web

pages and sites changes frequently.

2. The Methods that Filtering Companies Use to

Compile Category Lists

While the way in which filtering programs operate is

coneeptually straightforward—by comparing a re-

quested URL to a previously compiled list of URLs and

blocking access to the content at that URL if it appears

on the list—accurately compiling and categorizing

URLs to form the category lists is a more complex pro-

cess that is impossible to conduct with any high degree

of accuracy. The specific methods that filtering

Indeed, we granted leave for N2H2’s counsel to intervene in

order to object to testimony that would potentially reveal N2H2’s

trade secrets, which he did on several occasions.

54a

software companies use to compile and categorize

control lists are, like the lists themselves, proprietary

information. We will therefore set forth only general

information on the various types of methods that all

filtering companies deposed in this case use, and the

sources of error that are at once inherent in those

methods and unavoidable given the current architec-

ture of the Internet and the current state of the art in

automated classification systems. We base our under-

standing of these methods largely on the detailed

testimony and expert report of Dr. Geoffrey Nunberg,

which we credit. The plaintiffs offered, and the Court

qualified, Nunberg as an expert witness on automated

classification systems.

When compiling and categorizing URLs for their

category lists, filtering software companies go through

two distinct phases. First, they must collect or “har-

vest” the relevant URLs from the vast number of sites

that exist on the Web. Second, they must sort through

the URLs they have collected to determine under

which of the company’s self-defined categories (if any),

they should be classified. These tasks necessarily re-

sult in a tradeoff between overblocking (i.e., the block-

ing of content that does not meet the category

definitions established by CIPA or by the filtering

software companies), and underblocking (i.e., leaving off

7 Geoffrey Nunberg (Ph.D., Linguistics, C.U.N.Y.1977) is a

researcher at the Center for the Study of Language and Infor-

mation at Stanford University and a Consulting Full Professor of

Linguistics at Stanford University. Until 2001, he was also a

principal scientist at the Xerox Palo Alto Research Center. His

research centers on automated classification systems, with a focus

on classifying documents on the Web with respect to their linguis-

tie properties. He has published his research in numerous

professional journals, including peer-reviewed journals.

55a

of a control list a URL that contains content that would

meet the category definitions defined by CIPA or the

filtering software companies).

a. The “Harvesting” Phase

Filtering software companies, given their limited

resources, do not attempt to index or classify all of the

billions of pages that exist on the Web. Instead, the set

of pages that they attempt to examine and classify is

restricted to a small portion of the Web. The companies

use a variety of automated and manual methods to

identify a universe of Web sites and pages to “harvest”

for classification. These methods include: entering

certain key words into search engines; following links

from a variety of online directories (e.g., generalized

directories like Yahoo or various specialized directories,

such as those that provide links to sexually explicit

content); reviewing lists of newly-registered domain

names; buying or licensing lists of URLs from third

parties; “mining” access logs maintained by their cus-

tomers; and reviewing other submissions from cus-

tomers and the public. The goal of each of these

methods is to identify as many URLs as possible that

are likely to contain content that falls within the

filtering companies’ category definitions.

The first method, entering certain keywords into

commercial search engines, suffers from several limita-

tions. First, the Web pages that may be “harvested”

through this method are limited to those pages that

search engines have already identified. However, as

noted above, a substantial portion of the Web is not

even theoretically indexable (because it is not linked to

by any previously known page), and only approximately

50% of the pages that are theoretically indexable have

56a

actually been indexed by search engines. We are

satisfied that the remainder of the indexable Web, and

the vast “Deep Web,” which cannot currently be in-

dexed, includes materials that meet CIPA’s categories

of visual depictions that are obscene, child porno-

graphy, and harmful to minors. These portions of the

Web cannot presently be harvested through the

methods that filtering software companies use (except

through reporting by customers or by observing users’

log files), because they are not linked to other known

pages. A-user can, however, gain access to a Web site

in the unindexed Web or the Deep Web if the Web

site’s proprietor or some other third party informs the

user of the site’s URL. Some Web sites, for example,

send out mass email advertisements containing the

site’s URL, the spamming process we have described

above.

Second, the search engines that software companies

use for harvesting are able to search text only, not

images. This is of critical importance, because CIPA,

by its own terms, covers only “visual depictions.” 20

U.S.C. § 9134(f)(1)(A)G); 47 U.S.C. § 2540h) (5) (B) (i).

Image recognition technology is immature, ineffective,

and unlikely to improve substantially in the near future.

None of the filtering software companies deposed in

this case employs image recognition technology when

harvesting or categorizing URLs. Due to the reliance

on automated text analysis and the absence of image

recognition technology, a Web page with sexually

explicit images and no text cannot be harvested using a

search engine. This problem is complicated by the fact

that Web site publishers may use image files rather

than text to represent words, i.e., they may use a file

that computers understand to be a picture, like a

57a

photograph of a printed word, rather than regular text,

making automated review of their textual content

impossible. For example, if the Playboy Web site

displays its name using a logo rather than regular text,

a search engine would not see or recognize the Playboy

name in that logo.

In addition to collecting URLs through search

engines and Web directories (particularly those spe-

cializing in sexually explicit sites or other categories

relevant to one of the filtering companies’ category

definitions), and by mining user logs and collecting

URLs submitted by users, the filtering companies

expand their list of harvested URLs by using

“spidering” software that can “craw!” the lists of pages

produced by the previous four methods, following their

links downward to bring back the pages to which they

link (and the pages to which those pages link, and so on,

but usually down only a few levels). This spidering

software uses the same type of technology that com-

mercial Web search engines use.

While useful in expanding the number of relevant

URLs, the ability to retrieve additional pages through

this approach is limited by the architectural feature of

the Web that page-to-page links tend to converge

rather than diverge. That means that the more pages

from which one spiders downward through links, the

smaller the proportion of new sites one will uncover; if

spidering the links of 1000 sites retrieved through a

search engine or Web directory turns up 500 additional

distinct adult sites, spidering an additional 1000 sites

may turn up, for example, only 250 additional distinct

sites, and the proportion of new sites uncovered will

continue to diminish as more pages are spidered.

58a

These limitations on the technology used to harvest a

set of URLs for review will necessarily lead to sub-

stantial underblocking of material with respect to both

the category definitions employed by filtering software

companies and CIPA’s definitions of visual depictions

that are obscene, child pornography, or harmful to

minors.

b. The “Winnowing” or Categorization Phase

Once the URLs have been harvested, some filtering

software companies use automated key word analysis

tools to evaluate the content and/or features of Web

sites or pages accessed via a particular URL and to

tentatively prioritize or categorize them. This process

may be characterized as “winnowing” the harvested

URLs. Automated systems currently used by filtering

software vendors to prioritize, and to categorize or

tentatively categorize the content and/or features of a

Web site or page accessed via a particular URL operate

by means of (1) simple key word searching, and (2) the

use of statistical algorithms that rely on the frequency

and structure of various linguistic features in a Web

page’s text. The automated systems used to categorize

pages do not include image recognition technology. All

of the filtering companies deposed in the case also

employ human review of some or all collected Web

pages at some point during the process of categorizing

Web pages. As with the harvesting process, each

technique employed in the winnowing process is subject

to limitations that can result in both overblocking and

underblocking.

First, simple key-word-based filters are subject to

the obvious limitation that no string of words can

identify all sites that contain sexually explicit content,

59a

and most strings of words are likely to appear in Web

sites that are not properly classified as containing sexu-

ally explicit content. As noted above, filtering software

companies also use more sophisticated automated

classification systems for the statistical classification of

texts. These systems assign weights to words or other

textual features and use algorithms to determine

whether a text belongs to a certain category. These

algorithms sometimes make reference to the position of

a word within a text or its relative proximity to other

words. The weights are usually determined by machine

learning methods (often described as “artificial intel-

ligence”). In this procedure, which resembles an

automated form of trial and error, a system is given a

“training set” consisting of documents preclassified into

two or more groups, along with a set of features that

might be potentially useful in classifying the sets. The

system then “learns” rules that assign weights to those

features according to how well they work in classi-

fication, and assigns each new document to a category

with a certain probability.

Notwithstanding their “artificial intelligence”

description, automated text classification systems are

unable to grasp many distinctions between types of

content that would be obvious to a human. And of

critical importance, no presently conceivable technology

can make the judgments necessary to determine

whether a visual depiction fits the legal definitions of

obscenity, child pornography, or harmful to minors.

Finally, all the filtering software companies deposed

in this case use some form of human review in their

process of winnowing and categorizing Web pages,

although one company admitted to categorizing some

Web pages without any human review. SmartFilter

60a

states that “the final categorization of every Web site is

done by a human reviewer.” Another filtering company

asserts that of the 10,000 to 30,000 Web pages that

enter the “work queue” to be categorized each day, two

to three percent of those are automatically categorized

by their PornByRef system (which only applies to

materials classified in the pornography category), and

the remainder are categorized by human review.

SurfControl also states that no URL is ever added to

its database without human review.

Human review of Web pages has the advantage of

allowing more nuanced, if not more accurate, inter-

pretations than automated classification systems are

capable of making, but suffers from its own sources of

error. The filtering software companies involved here

have limited staff, of between eight and a few dozen

people, available for hand reviewing Web pages. The

reviewers that are employed by these companies base

their categorization decisions on both the text and the

visual depictions that appear on the sites or pages they

are assigned to review. Human reviewers generally

focus on English language Web sites, and are generally

not required to be multi-lingual.

Given the speed at which human reviewers must

work to keep up with even a fraction of the approxi-

mately 1.5 million pages added to the publicly indexable

Web each day, human error is inevitable. Errors are

likely to result from boredom or lack of attentiveness,

overzealousness, or a desire to “err on the side of

caution” by screening out material that might be offen-

sive to some customers, even if it does not fit within any

of the company’s category definitions. None of the

filtering companies trains its reviewers in the legal

definitions concerning what is obscene, child porno-

6la

graphy, or harmful to minors, and none instructs

reviewers to take community standards into account

when making categorization decisions.

Perhaps because of limitations on the number of

human reviewers and because of the large number of

new pages that are added to the Web every day,

filtering companies also widely engage in the practice of

categorizing entire Web sites at the “root URL,” rather

than engaging in a more fine-grained analysis of the

individual pages within a Web site. For example, the

filtering software companies deposed in this case all

categorize the entire Playboy Web site as Adult, Sex-

ually Explicit, or Pornography. They do not differenti-

ate between pages within the site containing sexually

explicit images or text, and for example, pages con-

taining no sexually explicit content, such as the text of

interviews of celebrities or politicians. If the “root” or

“top-level” URL of a Web site is given a category tag,

then access to all content on that Web site will be

blocked if the assigned category is enabled by a

customer.

In some cases, whole Web sites are blocked because

the filtering companies focus only on the content of the

home page that is accessed by entering the root URL.

Entire Web sites containing multiple Web pages are

commonly categorized without human review of each

individual page on that site. Web sites that may con-

tain multiple Web pages and that require authentica-

tion or payment for access are commonly categorized

based solely on a human reviewer’s evaluation of the

pages that may be viewed prior to reaching the

authentication or payment page.

Because there may be hundreds or thousands of

pages under a root URL, filtering companies make it

62a

their primary mission to categorize the root URL, and

categorize subsidiary pages if the need arises or if there

is time. This form of overblocking is called “inheri-

tance,” because lower-level pages inherit the categori-

zation of the root URL without regard to their specific

content. In some cases, “reverse inheritance” also

occurs, i.e., parent sites inherit the classification of

pages in a lower level of the site. This might happen

when pages with sexual content appear in a Web site

that is devoted primarily to non-sexual content. For

example, N2H2’s Bess filtering product classifies every

page in the Salon.com Web site, which contains a wide

range of news and cultural commentary, as “Sex,

Prof ity,“ based on the fact that the site includes a

regular column that deals with sexual issues.

Blocking by both domain name and IP address is

another practice in which filtering companies engage

that is a function both of the architecture of the Web

and of the exigencies of dealing with the rapidly

expanding number of Web pages. The category lists

maintained by filtering software companies can include

URLs in either their human-readable domain name

address form, their numeric IP address form, or both.

Through “virtual hosting” services, hundreds of thou-

sands of Web sites with distinct domain names may

share a single numeric IP address. To the extent that

filtering companies block the IP addresses of virtual

hosting services, they will necessarily block a sub-

stantial amount of content without reviewing it, and

will likely overblock a substantial amount of content.

Another technique that filtering companies use in

order to deal with a structural feature of the Internet is

blocking the root level URLs of so-called “loophole”

Web sites. These are Web sites that provide access to a

63a

particular Web page, but display in the user’s browser

a URL that is different from the URL with which the

particular page is usually associated. Because of this

feature, they provide a “loophole” that can be used to

get around filtering software, i.e., they display a URL

that is different from the one that appears on the

filtering company’s control list. “Loophole” Web sites

include caches of Web pages that have been removed

from their original location, “anonymizer” sites, and

translation sites.

Caches are archived copies that some search engines,

such as Google, keep of the Web pages they index. The

cached copy stored by Google will have a URL that is

different from the original URL. Because Web sites

often change rapidly, caches are the only way to access

pages that have been taken down, revised, or have

changed their URLs for some reason. For example, a

magazine might place its current stories under a given

URL, and replace them monthly with new stories. Ifa

user wanted to find an article published six months ago,

he or she would be unable to access it if not for Google’s

cached version.

Some sites on the Web serve as a proxy or inter-

mediary between a user and another Web page. When

using a proxy server, a user does not access the page

from its original URL, but rather from the URL of the

proxy server. One type of proxy service is an

“anonymizer.” Users may access Web sites indirectly

via an anonymizer when they do not want the Web site

they are visiting to be able to determine the IP address

from which they are accessing the site, or to leave

64a

“cookies” on their browser.“ Some proxy servers can

be used to attempt to translate Web page content from

one language to another. Rather tian directly access-

ing the original Web page in its original language, users

ean instead indirectly access the page via a proxy

server offering translation features.

As noted above, filtering companies often block

loophole sites, such as caches, anonymizers, and transla-

tion sites. The practice of blocking loophole sites nec-

essarily results in a significant amount of overblocking,

because the vast majority of the pages that are cached,

for example, do not contain content that would match a

filtering company’s category definitions. Filters that do

not block these loophole sites, however, may enable

users to access any URL on the Web via the loophole

site, thus resulting in substantial underblocking.

c. The Process for “Re-Reviewing”

Web Pages After Their Initial Categorization

Most filtering software companies do not engage in

subsequent reviews of categorized sites or pages on a

scheduled basis. Priority is placed on reviewing and

categorizing new sites and pages, rather than on re-

reviewing already categorized sites and pages. Typi-

cally, a filtering software vendor’s previous categoriza-

tion of a Web site is not re-reviewed for accuracy when

new pages are added to the Web site. To the extent the

Web site was previously categorized as a whole, the

8 A “cookie” is “a small file or part of a file stored on a World

Wide Web user’s computer, created and subsequently read by a

Web site server, and containing personal information (as a user

identification code, customized preferences, or a record of pages

visited).” Merriam Webster’s Collegiate Dictionary, available at

http:/hwww.m-w.com/dictionary.htm.

65a

new pages added to the site usually share the cate-

gorization assigned by the blocking product vendor.

This necessarily results in both over- and under-

blocking, because, as noted above, the content of Web

pages and Web sites changes relatively rapidly.

In addition to the content on Web sites or pages

changing rapidly, Web sites themselves may disappear

and be replaced by sites with entirely different content.

If an IP address associated with a particular Web site is

blocked under a particular category and the Web site

goes out of existence, then the IP address likely would

be reassigned to a different Web site, either by an

Internet service provider or by a registration organi-

zation, such as the American Registry for Internet

Numbers, see http:/;www.arin.net. In that case, the site

that received the reassigned IP address would likely be

miscategorized. Because filtering companies do not

engage in systematic re-review of their category lists,

such a site would likely remain miscategorized unless

someone submitted it to the filtering company for re-

review, increasing the incidence of over- and under-

blocking.

This failure to re-review Web pages primarily in-

creases a filtering company’s rate of overblocking.

However, if a filtering company does not re-review

Web pages after it determines that they do not fall into

any of its blocking categories, then that would result in

underblocking (because, for example, a page might add

sexually explicit content).

3. The Inherent Tradeoff Between Overblocking and

Underblocking

There is an inherent tradeoff between any filter’s

rate of overblocking (which information scientists also

66a

call “precision”) and its rate of underblocking (which is

also referred to as recall“). The rate of overblocking or

precision is measured by the proportion of the things a

classification system assigns to a certain category that

are appropriately classified. The plaintiffs’ expert, Dr.

Nunberg, provided the hypothetical example of a

classification system that is asked to pick out pictures

of dogs from a database consisting of 1000 pictures of

animals, of which 80 were actually dogs. If it returned

100 hits, of which 80 were in fact pictures of dogs, and

the remaining 20 were pictures of cats, horses, and

deer, we would say that the system identified dog pic-

tures with a precision of 80%. This would be analogous

to a filter that overblocked at a rate of 20%.

The recall measure involves determining what pro-

portion of the actual members of a category the classi-

fication system has been able to identify. For example,

if the hypothetical animal-picture database contained a

total of 200 pictures of dogs, and the system identified

80 of them and failed to identify 120, it would have

performed with a recall of 40%. This would be analo-

gous to a filter that underblocked 60% of the material in

a category.

In automated classification systems, there is always a

tradeoff between precision and recall. In the animal-

picture example, the recall could be improved by using

a looser set of criteria to identify the dog pictures in the

set, such as any animal with four legs, and all the dogs

would be identified, but cats and other animals would

also be included, with a resulting loss of precision. The

same tradeoff exists between rates of overblocking and

underblocking in filtering systems that use automated

classification systems. For example, an automated

system that classifies any Web page that contains the

67a

word “sex” as sexually explicit will underblock much

less, but overblock muck. more, than a system that clas-

sifies any Web page containing the phrase “free pic-

tures of people having sex” as sexually explicit.

This tradeoff between overblocking and under-

blocking also applies not just to automated classification

systems, but also to filters that use only hurnan review.

Given the approximately two billion pages that exist on

the Web, the 1.5 million new pages that are added daily,

and the rate at which content on existing pages

changes, if a filtering company blocks only those Web

pages that have been reviewed by humans, it will be

impossible, as a practical matter, to avoid vast amounts

of underblocking. Techniques used by human re-

viewers such as blocking at the IP address level,

domain name level, or directory level reduce the rates

of underblocking, but necessarily increase the rates of

overblocking, as discussed above.

To use a simple example, it would be easy to design a

filter intended to block sexually explicit speech that

completely avoids overblocking. Such a filter would

have only a single sexually explicit Web site on its

control list, which could be reviewed daily to ensure

that its content does not change. While there would be

no overblocking problem with such a filter, such a filter

would have a severe underblocking problem, as it would

fail to block all the sexually explicit speech on the Web

other than the one site on its control list. Similarly, it

would also be easy to design a filter intended to block

sexually explicit speech that completely avoids under-

blocking. Such a filter would operate by permitting

users to view only a single Web site, e.g., the Sesame

Street Web site. While there would be no under-

blocking problem with such a filter, it would have a

68a

severe overblocking problem, as it would block access

to millions of non-sexually explicit sites on the Web

other than the Sesame Street site.

While it is thus quite simple to design a filter that

does not overblock, and equally simple to design a filter

that does not underblock, it is currently impossible,

given the Internet’s size, rate of growth, rate of change,

and architecture, and given the state of the art of

automated classification systems, to develop a filter

that neither underblocks nor overblocks a substantial

amount of speech. The more effective a filter is at

blocking Web sites in a given category, the more the

filter will necessarily overblock. Any filter that is

reasonably effective in preventing users from accessing

sexually explicit content on the Web will necessarily

block substantial amounts of non-sexually explicit

speech.

4. Attempts to Quantify Filtering Programs’ Rates

of Over-and Underblocking

The government presented three studies, two from

expert witnesses, and one from a librarian fact witness

who conducted a study using Internet use logs from his

own library, that attempt to quantify the over- and

underblocking rates of five different filtering programs.

The plaintiffs presented one expert witness who at-

tempted to quantify the rates of over- and under-

blocking for various programs. Each of these attempts

to quantify rates of over- and underblocking suffers

from various methodological flaws.

The fundamental problem with calculating over- and

underblocking rates is selecting a universe of Web sites

or Web pages to serve as the set to be tested. The

studies that the parties submitted in this case took two

69a

different approaches to this problem. Two of the

studies, one prepared by the plaintiffs’ expert witness

Chris Hunter, a graduate student at the University of

Pennsylvania, and the other prepared by the defen-

dants’ expert, Chris Lemmons of eTesting Labora-

tories, in Research Triangle Park, North Carolina,

approached this problem by compiling two separate

lists of Web sites, one of URLs that they deemed

should be blocked according to the filters’ criteria, and

another of URLs that they deemed should not be

blocked according to the filters’ criteria. They compiled

these lists by choosing Web sites from the results of

certain key word searches.“ The problem with this

Hunter drew three different “samples” for his test. The first

consisted of “50 randomly generated Web pages from the Web-

crawler search engine.” The “second sample of 50 Web pages was

drawn from searches for the terms ‘yahoo, warez, hotmail, sex, and

MP3,’ using the AltaVista.com search engine.” And the “final

sample of 100 Web sites was drawn from the sites of organizations

who filed amicus briefs in support of the ACLU’s challenges to the

Community [sic] Decency Act (CDA) and COPA [the Children’s

Online Protection Act], and from Internet portals, political Web

sites, feminist Web sites, hate speech sites, gambling sites,

religious sites, gay pride/homosexual sites, alcohol, tobacco, and

drug sites, pornography sites, new sites, violent game sites, safe

sex sites, and pro and anti-abortion sites listed on the popular Web

directory, Yahoo.com.”

Lemmons testified that he compiled the list of sexually explicit

sites that should have been blocked by entering the terms “free

adult sex, anal sex, oral sex, fisting lesbians, gay sex, interracial

sex, big tits, blow job, shaved pussy, and bondage” into the Google

search engine and then “surfing” through links from pages

generated by the list of sites that the search engine returned.

Using this method, he compiled a list of 197 sites that he deter-

mined should be blocked according to the filtering programs’

category definitions. Lemmons also attempted to compile a list of

“sensitive” Web sites that, although they should not have been

70a

selection method is that it is neither random, nor does it

necessarily approximate the universe of Web pages

that library patrons visit.

The two other studies, one by David Biek, head

librarian at the Tacoma Public Library’s main branch,

and one by Cory Finnell of Certus Consulting Group, of

Seattle, Washington, chose actual logs of Web pages

visited by library patrons during specific time periods

as the universe of Web pages to analyze. This method,

while surely not as accurate as a truly random sample

of the indexed Web would be (assuming it would be

possible to take such a sample), has the virtue of using

the actual Web sites that library patrons visited during

a specific period. Because library patrons selected the

universe of Web sites that Biek and Finnell's studies

analyzed, this removes the possibility of bias resulting

from the study author’s selection of the universe of

sites to be reviewed. We find that the Lemmons and

Hunter studies are of little probative value because of

the methodology used to select the sample universe of

Web sites to be tested. We will therefore focus on the

studies conducted by Finnell and Biek in trying to as-

certain estimates of the rates of over- and under-

blocking that takes place when filters are used in public

libraries.

blocked according to the filtering programs’ category definitions,

might have been mistakenly blocked. In order to do this, he used

the same method of entering terms into the Google search engine

and surfing through the results. He used the following terms to

compile this list: “breast feeding, bondages, fetishes, ebony, gay

issues, women’s health, lesbian, homosexual, vagina, vaginal dry-

ness, pain, anal cancer, teen issues, safe sex, penis, preg ant, inter

racial, sex education, penis enlargement, breast enlargement,. . .

and shave.”

71a

The government hired expert witness Cory Finnell

to study the Internet logs compiled by the public

libraries systems in Tacoma, Washington; Westerville,

Ohio; and Greenville, South Carolina. Each of these

libraries uses filtering software that keeps a log of

information about individual Web site requests made

by library patrons. Finnell, whose consulting firm

specializes in data analysis, has substantial experience

evaluating Internet access logs generated on net-

worked systems. He spent more than a year develop-

ing a reporting tool for N2H2, and, in the course of that

work, acquired a familiarity with the design and

operation of Internet filtering products.

The Tacoma library uses Cyber Patrol filtering

software, and logs information only on sites that were

blocked. Finnell worked from a list of all sites that were

blocked in the Tacoma public library in the month of

August 2001. The Westerville library uses the Web-

sense filtering product, and logs information on both

blocked sites and non-blocked sites. When the logs

reach a certain size, they are overwritten by new usage

logs. Because of this overwriting feature, logs were

available to Finnell only for the relatively short period

from October 1, 2001 to October 3, 2001. The Greenville

library uses N2H2’s filtering product and logs both

blocken sites and sites that patrons accessed. The logs

contain more than 500,000 records per day. Because of

the volume of the records, Finnell restricted his

— to the period from August 2, 2001 to August 15,

Finnell calculated an overblocking rate for each of

the three libraries by examining the host Web site

containing each of the blocked pages. He did not

employ a sampling technique, but instead examined

72a

each blocked Web site. If the contents of a host Web

site or the pages within the Web site were consistent

with the filtering product’s definition of the category

under which the site was blocked, Finnell considered it

to be an accurate block. Finnell and three others, two

of whom were ten ary employees, examined the

Web sites to determine whether they were consistent

with the filtering companies’ category definitions.

Their review was, of course, necessarily limited by:

(1) the clarity of the filtering companies’ category defi-

nitions; (2) Finnell’s and his employees’ interpretations

of the definitions; and (3) human error. The study’s

reliability is also undercut by the fact that Finnell failed

to archive the blocked Web pages as they existed either

at the point that a patron in one of the three libraries

was denied access or when Finnell and his team

reviewed the pages. It is therefore impossible for any-

one to check the accuracy and consistency of Finnell’s

review team, or to know whether the pages contained

the same content when the block occurred as they did

when Finnell’s team reviewed them. This is a key flaw,

because the results of the study depend on individual

determinations as to overblocking and underblocking,

in which Finnell and his team were required to compare

what they saw on the Web pages that they reviewed

with standard definitions provided by the filtering

company.

Tacoma library’s Cyber Patrol software blocked 836

unique Web sites during the month of August. Finnell

determined that 783 of those blocks were accurate and

that 53 were inaccurate.” The error rate for Cyber

1© If separate patrons attempted to reach the same Web site, or

one or more patrons attempted to access more than one Page On 8

single Web site, Finnell counted these attempts as a single

73a

Patrol was therefore estimated to be 6.34%. an

34%, and th

—— rate was estimated with 95% confidence to

within the range of 4.69% to 7.99%." Finnell and his

team reviewed 185 unique Web sites that were blocked

by Westerville Library’s Websense filter during the

logged period and determined that 158 of them were

accurate and that 27 of them were inaccurate. He

therefore estimated the Websense filter’s overblocking

— at 14.59% with a 95% confidence interval of 9.51%

9.68%. Additionally, Finnell examined 1,674 unique

Web sites that were blocked by the Greenville Lib-

— N2H2 filter during the relevant period and

é termined that 1,520 were accurate and that 87 were

rr This yields an estimated overblocking rate

41 and a 95% confidence interval of 4.33% to

that the filters were operating, and ma

' , y have bee

deterred from attempting to access Web sites that they

perceived to be “borderline” sites, i. e., those that may

or may not have been appropriately filtered according

to the filtering companies’ category definitions. Second,

74a

i ir cross-examination of Finnell, the plaintiffs

223 screen shots of a number of Web sites that,

according to Finnell, had been appropriately blocked,

but that Finnell admitted contained only benign

materials. Finnell's explanation was that the Web sites

must have changed between the time when he con-

dueted the study and the time of the trial, but because

he did not archive the images as they existed when his

team reviewed them for the study, there is no way to

verify this. Third, because of the way in which Finnell

counted blocked Web sites—i.e., if separate patrons

attempted to reach the same Web site, or one or more

patrons attempted to access more than one page on a

single Web site, Finnell counted these attempts as a

single block, see supra note 10—his results necessarily

understate the number of times that patrons were

-erroneously denied access to information.

At all events, there is no doubt that Finnell's esti-

mated rates of overblocking, which are based on the

filtering companies’ own category definitions, signifi-

cantly understate the rate of overblocking with respect

to CIPA’s category definitions for filtering for adults.

The filters used in the Tacoma, Westerville, and Green-

ville libraries were configured to block, among other

things, images of full nudity and sexually explicit mate-

rials. There is no dispute, however, that these catego-

ries are far broader than CIPA’s categories of visual

depictions that are obscene, or child pornography, the

two categories of material that libraries subject to

CIPA must certify that they filter during adults’ use of

the Internet.

Finnell’s study also calculated underblocking rates

with respect to the Westerville and Greenville Librar-

ies (both of which logged not only their blocked sites,

75a

but all sites visited by their patrons), by taking random

samples of URLs from the list of sites that were not

blocked. The study used a sample of 159 sites that were

accessed by Westerville patrons and determined that

only one of them should have been blocked under the

soft ware's category definitions, yielding an under-

blocking rate of 0.6%. Given the size of the sample, the

95% confidence interval is 0% to 1.86%. The study

examined a sample of 254 Web sites accessed by pa-

trons in Greenville and found that three of them should

have been blocked under the filtering software’s cate-

gory definitions. This results in an estimated under-

blocking rate of 1.2% with a 95% confidence interval

ranging from 0% to 2.51%.

We do not credit Finnell’s estimates of the rates of

underblocking in the Westerville and Greenville public

libraries for several reasons. First, Finnell’s estimates

likely understate the actual rate of underblocking be-

cause patrons, who knew that filtering programs were

operating in the Greenville and Westerville Libraries,

may have refrained from attempting to access sites

with sexually explicit materials, or other contents that

they knew would probably meet a filtering program’s

blocked categories. Second, and most importantly, we

think that the formula that Finnell used to calculate the

rate of underblocking in these two libraries is not as

meaningful as the formula that information scientists

typically use to calculate a rate of recall, which we

describe above in Subsection II.E.3. As Dr. Nunberg

explained, the standard method that information

scientists use to calculate a rate of recall is to sort a set

of items into two groups, those that fall into a particular

category (e.g., those that should have been blocked by a

filter) and those that do not. The rate of recall is then

76a

ividi that the

leulated by dividing the number of items

— correctly identified as belonging to the category

by the total number of items in the category.

base that

In the example above, we discussed a data

contained 1000 photographs. Assume that 200 of these

i at a rate of 60%. To calculate the recall rate of

~~ in the Westerville and Greenville ——

libraries in accordance with the standard meth

described above, Finnell should have taken a sample 0

sites from the libraries’ Internet use logs (including

both sites that were blocked and sites that were =

and divided the number of sites in the sample that ?

filter incorrectly failed to block by the total number a

sites in the sample that should have been block ,

What Finnell did instead was to take a sample of —

that were not blocked, and divide the total number

sites in this sample by the number of sites in the sample

that should have been blocked. This made the denomi-

nator that Finnell used much larger than it would have

been had he used the standard method for calculating

recall, consequently making the underblocking rate that

he calculated much lower than it would have been

under the standard method.”

i j random

29 inustrate the two different methods, consider a

— 1010 web sites taken from a library's Internet use log, 10

of which fall within the category that a filter is intended to bit

(eg, pornography), and suppose that the filter incorrectly it Plock

block 2 of the 10 sites that it should have blocked and

any sites that should not have been blocked. The standard method

77a

Moreover, despite the relatively low rates of under-

blocking that Finnell's study found, librarians from

several of the libraries proffered by defendants that use

blocking products, including Greenville, Tacoma, and

Westerville, testified that there are instances of under-

blocking in their libraries. No quantitative evidence

was presented comparing the effectiveness of filters

and other alternative methods used by libraries to

prevent patrons from accessing visual depictions that

are obscene, child pornography, or in the case of minors,

harmful to minors.

Biek undertook a similar study of the overblocking

rates that result from the Tacoma Library’s use of the

Cyber Patrol software. He began with the 3,733 indivi-

dual blocks that occurred in the Tacoma Library in

October 2000 and drew from this data set a random

sample of 786 URLs. He calculated two rates of over-

blocking, one with respect to the Tacoma Library’s

policy on Internet use—that the pictorial content of the

site may not include “graphic materials depicting full

nudity and sexual acts which are portrayed obviously

and exclusively for sensational or pornographic pur-

poses”—and the other with respect to Cyber Patrol’s

own category definitions. He estimated that Cyber

Patrol overblocked 4% of all Web pages in October 2000

with respect to the definitions of the Tacoma Library’s

of quantifying the rate of underblocking would divide the number

of sites in the sample that the filter incorrectly failed to block by

the number of sites in the sample that the filter should have

blocked, yielding an underblocking rate in this example of 20%.

Finnell’s study, however, calculated the underblocking rate by

dividing the number of sites that the filter incorrectly failed to

block by the total number of sites in the sample that were not

blocked (whether correctly or incorrectly) yielding an under-

blocking rate in this example of only 2%.

78a

Internet Policy and 2% of all pages with respect to

Cyber Patrol’s own category definitions.”

It is difficult to determine how reliable Biek’s con-

clusions are, because he did not keep records of the raw

data that he used in his study; nor did he archive

images of the Web pages as they looked when he made

the determination whether they were properly classi-

fied by the Cyber Patrol program. Without this infor-

mation, it is impossible to verify his conclusions (or to

undermine them). And Biek’s study certainly under-

states Cyber Patrol’s overblocking rate for some of the

same reasons that Finnell’s study likely understates the

true rates of overblocking used in the libraries that he

studied.

We also note that Finnell’s study, which analyzed a

set of Internet logs from the Tacoma Library during

which the same filtering program was operating with

the same set of blocking categories enabled, found a

significantly higher rate of overblocking than the Biek

study did. Biek found a rate of overblocking of

approximately 2% while the Finnell study estimated a

6.34% rate of overblocking. At all events, the category

definitions employed by CIPA, at least with respect to

adult use—visual depictions that are obscene or child

pornography—are narrower than the materials prohib-

ited by the Tacoma Library policy, and therefore Biek’s

study understates the rate of overblocking with respect

to CIPA’s definitions for adults.

In sum, we think that Finnell’s study, while we do

not credit its estimates of underblocking, is useful

because it states lower bounds with respect to the rates

13 According to Biek, the sample size that he used yielded a 95%

confidence interval of plus or minus 3.11%.

79a

of overblocking that occurred when the Cyber Patrol,

Websense, and N2H2 filters were operating in public

libraries. While these rates are substantial—between

nearly 6% and 15%—we think, for the reasons stated

above, that they greatly understate the actual rates of

overblocking that occurs, and therefore cannot be

considered as anything more than minimum estimates

of the rates of overblocking that happens in all filtering

programs.

5. Methods of Obtaining Examples of Erroneously

Blocked Web Sites

The plaintiffs assembled a list of several thousand

Web sites that they contend were, at the time of the

study, likely to have been erroneously blocked by one

or more of four major commercial filtering programs:

SurfControl Cyber Patrol 6.0.1.47, N2H2 Internet

Filtering 2.0, Secure Computing SmartFilter 3.0.0.01,

and Websense Enterprise 4.3.0. They compiled this list

using a two- step process. First, Benjamin Edelman, an

expert witness who testified before us, compiled a list

of more than 500,000 URLs and devised a program to

feed them through all four filtering programs in order

to compile a list of URLs that might have been

erroneously blocked by one or more of the programs.“

Second, Edelman forwarded subsets of the list that he

compiled to librarians and professors of library science

whom the plaintiffs had hired to review the blocked

sites for suitability in the public library context.

Edelman is a Harvard University student and a systems

administrator and multimedia specialist at the Berkman Center for

Internet and Society at Harvard Law School. Despite Edelman’s

young age, he has been doing consulting work on Internet-related

issues for nine years, since he was in junior high school.

80a

Edelman assembled the list of URLs by compiling

Web pages that were blocked by the following cate-

gories in the four programs: Cyber Patrol: Adult/

Sexually Explicit; N2H2: Adults Only, Nudity, Porno-

graphy, and Sex, with “exceptions” engaged in the

categories of Education, For Kids, History, Medical,

Moderated, and Text/Spoken Only; SmartFilter: Sex,

Nudity, Mature, and Extreme; Websense: Adult

Content, Nudity, and Sex.

Edelman then assembled a database of Web sites for

possible testing. He derived this list by automatically

compiling URLs from the Yahoo index of Web sites,

taking them from categories from the Yahoo index that

differed significantly from the classifications that he

had enabled in each of the blocking programs (taking,

for example, Web sites from Yahoo’s “Government”

category). He then expanded this list by entering

URLs taken from the Yahoo index into the Google

search engine’s “related” search function, which pro-

vides the user with a list of similar sites. Edelman also

included and excluded specific Web sites at the request

of the plaintiffs’ counsel.

Taking the list of more than 500,000 URLs that he

had compiled, Edelman used an automated system that

he had developed to test whether particular URLs

were blocked by each of the four filtering programs.

This testing took place between February and October

2001. He recorded the specific dates on which par-

ticular sites were blocked by particular programs, and,

using commercial archiving software, archived the

contents of the home page of the blocked Web sites (and

in some instances the pages linked to from the home

81a

page) as it existed when it was blocked.” Through this

process, Edelman, whose testimony we credit, compiled

a list of 6,777 URLs that were blocked by one or more

of the four programs. Because these sites were chosen

from categories from the Yahoo directory that were

unrelated to the filtering categories that were enabled

during the test (i.e., “Government” vs. “Nudity”), he

reasoned that they were likely erroneously blocked. As

explained in the margin, Edelman repeated his testing

and discovered that Cyber Patrol had unblocked most

of the pages on the list of 6,777 after he had published

the list on his Web site. His records indicate that an

employee of SurfControl (the company that produces

Cyber Patrol software) accessed his site and presuma-

bly checked out the URLs on the list, thus confirming

Edelman’s judgment that the majority of URLs on the

list were erroneously blocked.“

15 The archiving process in some cases took up to 48 hours from

when the page was blocked.

16 In October 2001, Edelman published the results of his initial

testing on his Web site. In February and March 2002 he repeated

his testing of the 6,777 URLs originally found to be blocked by at

least one of the blocking products, in order to determine whether

and to what extent the blocking product vendors had corrected the

mistakes that he publicized. Of those URLs blocked by N2H2 in

the October 2001 testing, 55.10% remained blocked when tested by

Edelman in March 2002. Of those URLs blocked by Websense in

the October 2001 testing, 76.28% remained blocked when tested by

Edelman in February 2002. Of those URLs blocked by Surf-

Control’s Cyber Patrol product, only 7.16% remained blocked, i.e.,

Cyber Patrol had unblocked almost 93% of the Web pages origi-

nally blocked. Because the results posted to his Web site were

accessed by an employee of SurfControl (as evidenced by Edel-

man’s records of who was accessing his Web site), we infer that

Cyber Patrol had determined that 93% of all 6,777 pages, or 6,302

Web pages, were originally wrongly blocked by the product.

82a

man forwarded the list of blocked sites to Dr.

J oun Janes, an Assistant Professor in the —

School of the University of Washington who a —

testified at trial as an expert witness. J anes —

the sites that Edelman compiled to determine whe “4

they are consistent with library collection development,

i. e., whether they are sites to which a —

librarian would, consistent with professional stand g

direct a patron as a source of information.

Edelman forwarded Janes a list of 6,775 Web sites,

almost the entire list of blocked sites that he collected,

from which Janes took a random sample of 859 using

i i list of

17 Two other expert witnesses reviewed subsets of the

Web pages that Edelman compiled. Dr. Michael T. ge toe ay

of the Rare Book and Manuscript Library and of the — .

Electronie Text and Image at the University of Pennsy ——

reviewed a list of 204 sites that Edelman forwarded to him 1

to determine their appropriateness and usefulness in the rary

setting. Because the sites that Ryan reviewed were not — —

randomly (i. e., they were chosen by plaintiffs — — —

says little about the character of the set of 6,7 he lp —

Edelman compiled, or the total amount of overblocking by

filtering programs that Edelman used. ;

Anne Lipow, a practicing librarian for more than 30 years an

list of 204 URLs from the set that Edelman had — —

appropriateness for a library 's collection. She catego : si =

four different levels according to their appropriateness for * pu

library’s collection. Again, because these URLs were a poe

randomly, Lipow’s study is not particularly relevant ti )

set that Edelman compiled, or to the total amount of ov ocking

by the four filtering programs that Edelman used. oie

Although the methodology used to select the list of —

that was forwarded to Ryan and Lipow is problematic, Ryan’s

Lipow’s testimony established that many of the —— y

blocked sites that Edelman identified would be useful appro-

priate sources of information for library patrons.

— 83a

the SPSS statistical software package. Janes indicated

that he chose a sample size of 859 because it would yield

a 95% confidence interval of plus or minus 2.5%. Janes

recruited a group of 16 reviewers, most of whom were

current or former students at the University of

Washington’s Information School, to help him identify

which sites were appropriate for library use. We

describe the process that he used in the margin.” Due

18 All of the reviewers that Janes recruited had some relevant

experience in library reference services or library collection de-

velopment. Janes divided the reviewers into two groups, a group

of 11 less experienced reviewers, and a group of five more experi-

enced reviewers. Janes assigned the less experienced group to doa

first-round review with the purpose of identifying the most ob-

viously overblocked sites. The more experienced group was to

review the remaining sites (i.e., those that were not obviously

overblocked) and to make final decisions regarding these sites.

In the first round, each person evaluated two sets of around 80

sites, and each group was evaluated by two different people. Each

set of sites included the following instructions:

Look carefully at each of the Web sites on the list. Please make

a notation of any site that appears to meet any of the following

a. Contains information similar to that already found in

libraries,

or

b. Contains information a librarian would want in the library

if s/he had unlimited funds to purchase information and

unlimited shelf space,

or

e. You would be willing to refer a patron (of any age) to the

site if the patron appeared at a reference desk seeking infor-

mation about the subject of the site. For this last criterion, we

recognize that you might not refer a young child to a Calculus

site just because it would not be useful to that child, but you

should ignore that factor. Informational sites, such as a

84a

to the inability of a member of Janes’s review team to

complete the reviewing process, Janes had to cut 157

Web sites out of the sample, but because the Web sites

were randomly assigned to reviewers, it is unlikely that

these sites differed significantly from the rest of the

sample. That left the sample size at 699, which widened

the 95% confidence interval to plus or minus 2.8%.

Of the total 699 sites reviewed, Janes’s team con-

cluded that 165 of them, or 23.6% percent of the sample,

were not of any value in the library context (i. e., no

librarian would, consistent with professional standards,

refer a patron to these sites as a source of information).

They were unable to find 60 of the Web sites, or 8.6% of

the sample. Therefore, they concluded that the remain-

ing 474 Web sites, or 67.8% of the sample, were exam-

ples of overblocking with respect to materials that are

appropriate sources of information in public libraries.

Calculus site, should be noted. A site that is purely erotica

should not be noted.

Sites that received “Yes” votes from both reviewers were deter-

mined to be of sufficient interest in a library context and removed

from further analysis. Sites receiving one or two “No” votes would

go to the next round. In the first round, 243 sites received “Yes”

votes from both reviewers, while 456 sites received one or more

“No” votes or could not be found. These 456 sites were sent for-

ward to the second round of judging.

The instructions for the second-round reviewers were the same

as those given to the first-round reviewers, except that in section

c, the following sentence was added: “Sites that have a commercial

purpose should be included here if they might be of use or interest

to someone wishing to buy the product or service or doing research

on commercial behavior on the Internet, much as most libraries

include the Yellow Pages in their collections.” The second round of

review produced the following results: 60 sites could not be found

(due to broken links, 404 “not found” errors, domain for sale mes-

sages, etc.), 231 sites were judged “Yes,” and 165 judged “No.”

85a

Applying a 95% confidence interval of plus or mi

2.8%, the study concluded that we can be 95% ——

that the actual percentage of sites in the list of 6,775

sites that are appropriate for use in public libraries is

somewhere between 65.0% and 70.6%. In other words

we can be 95% certain that the actual number of sites

out of the 6,775 that Edelman forwarded to Janes that

are appropriate for use in public libraries (under

* standard) is somewhere between 4,403 and

The government raised some valid criticisms of

Janes’s methodology, attacking in particular the fact

that, while sites that received two “yes” votes in the

first round of voting were determined to be of sufficient

interest ina library context to be removed from further

analysis, sites receiving one or two “no” votes were

sent to the next round. The government also correctly

points out that results of Janes’s study can be general-

ized only to the population of 6,775 sites that Edelman

forwarded to Janes. Even taking these criticisms into

account, and discounting Janes’s numbers appropri-

ately, we credit Janes’s study as confirming that Edel-

man’s set of 6,775 Web sites contains at least a few

thousand URLs that were erroneously blocked by one

or more of the four filtering programs that he used

whether judged against CIPA’s definitions, the filters’

own category criteria, or against the standard that the

Janes study used. Edelman tested only 500,000 unique

URLs out of the 4000 times that many, or two billion,

that are estimated to exist in the indexable Web. Even

assuming that Edelman chose the URLs that were

most likely to be erroneously blocked by commercial

filtering programs, we conclude that many times the

number of pages that Edelman identified are errone-

86a

ously blocked by one or more of the filtering programs

that he tested.

Edelman’s and Janes’s studies provide numerous

specific examples of Web pages that were erroneously

blocked by one or more filtering programs. The Web

pages that were erroneously blocked by one or more of

the filtering programs do not fall into any neat patterns;

they range widely in subject matter, and it is difficult to

tell why they may have been overblocked. The list that

Edelman compiled, for example, contains Web pages

relating to religion, politics and government, health, ca-

reers, education, travel, sports, and many other topics.

In the next section, we provide examples from each of

these categories.

6. Examples of Erroneously Blocked Web Sites

Several of the erroneously blocked Web sites had

content relating to churches, religious orders, religious

charities, and religious fellowship organizations. These

included the following Web sites: the Knights of Colum-

bus Council 4828, a Catholic men’s group associated

with St. Patrick's Church in Fallon, Nevada, http://

msnhomepages.talkcity.com/SpiritSt/kofc4828, which

was blocked by Cyber Patrol in the “Adult/Sexually

Explicit” category; the Agape Church of Searcy,

Arkansas, http:/hwww.agapechurch.com, which was

blocked by Websense as “Adult Content”; the home

page of the Lesbian and Gay Havurah of the Long

Beach, California Jewish Community Center, http://

www.compupiz.com/gay/havurah.htm, which was

blocked by N2H2 as “Adults Only, Pornography,” by

Smartfilter as “Sex,” and by Websense as “Sex”; Or-

phanage Emmanuel, a Christian orphanage in Hondu-

ras that houses 225 children, http://home8.inet.tele.

87a

dk/rfb5y(2)27viva, which was blocked by Cyber Patrol

in the “Adult/Sexually Explicit” category; Vision Art

Online, which sells wooden wall hangings for the

home that contain prayers, passages from the Bible,

and images of the Star of David, http:/howw.

visionartonline.com, which was blocked in Websense’s

“Sex” category; and the home page of Tenzin Palmo, a

Buddhist nun, which contained a description of her

project to build a Buddhist nunnery and international

retreat center for women, http://www.tenzinpalmo.com,

which was categorized as “Nudity” by N2H2.

Several blocked sites also contained information

about governmental entities or specific political can-

didates, or contained political commentary. These

included: the Web site for Kelley Ross, a Libertarian

candidate for the California State Assembly, http

www friesian.com/ross/ca40, which N2H2 blocked as

“Nudity”; the Web site for Bob Coughlin, a town select-

man in Dedham, Massachusetts, http://;www.bob

coughlin.org, which was blocked under N2H2’s

“N udity” category; a list of Web sites containing infor-

mation about government and politics in Adams

County, Pennsylvania, http:/hvww.geocities.com/

adamscopa, which was blocked by Websense as “Sex”;

the Web site for Wisconsin Right to Life, http://

www.wrtl.org, which N2H2 blocked as “Nudity”; a Web

site that promotes federalism in Uganda, http://federo.

com, which N2H2 blocked as “Adults Only, Porno-

graphy”; “Fight the Death Penalty in the USA,” a

Danish Web site dedicated to criticizing the American

system of capital punishment, http:/hwww.fdp.dk, which

N2H2 blocked as “Pornography”; and Dumb Laws,” a

humor Web site that makes fun of outmoded laws,

88a

http:/hwww.dumblaws.com, which N2H2 blocked under

its “Sex” category.

Erroneously blocked Web sites relating to health

issues included the following: a guide to allergies,

http://www.x-sitez.com/allergy, which was categorized

as “Adults Only, Pornography” by N2H2; a health ques-

tion and answer site sponsored by Columbia Univer-

sity, http:/;www.goaskalice.com.columbia.edu, which

was blocked as “Sex” by N2H2, and as “Mature” by

Smartfilter; the Western Amputee Support Alliance

Home Page, http:/hwww.usinter.net/wasa, which was

blocked by N2H2 as “Pornography”; the Web site of the

Willis-Knighton Cancer Center, a Shreveport, Louisi-

ana cancer treatment facility, http://cancerftr.wkmce.

com, which was blocked by Websense under the “Sex”

category; and a site dealing with halitosis, http://

www.dreamcastle.com/tungs, which was blocked by

N2H2 as “Adults, Pornography,” by Smartfilter as

“Sex,” by Cyber Patrol as “Adult/Sexually Explicit,”

and by Websense as “Adult Content.”

The filtering programs also erroneously blocked

several Web sites having to do with education and

careers. The filtering programs blocked two sites that

provide information on home schooling. “Hom Edu

Station—the Internet Source for Home Education,”

http://www.perigee.net/memullen/homedu station/, was

categorized by Cyber Patrol as “Adult/Sexually

Explicit.” Smartfilter blocked “Apricot: A Web site

made by and for home schoolers,” http:// apricotpie.com,

as “Sex.” The programs also miscategorized several

career-related sites. “Social Work Search,” hitp:/howw.

socialworksearch.com/, is a directory for social workers

that Cyber Patrol placed in its “Adult/Sexually Ex-

plicit” category. The “Gay and Lesbian Chamber of

89a

Southern Nevada,” http:/;www.lambdalv.com, “a forum

for the business community to develop relationships

within the Las Vegas lesbian, gay, transsexual, and

bisexual community” was blocked by N2H2 as “Adults

Only, Pornography.” A site for aspiring dentists,

http:/howw.vvm.com/bond/home.htm, was blocked by

Cyber Patrol in its “Adult/Sexually Explicit” category.

The filtering programs erroneously blocked many

travel Web sites, including: the Web site for the Allen

Farmhouse Bed & Breakfast of Alleghany County,

North Carolina, http://planet-nc.com/Beth/index.html,

which Websense blocked as “Adult Content”; Odysseus

Gay Travel, a travel company serving gay men, http.

www.odyusa.com, which N2H2 categorized as “Adults

Only, Pornography”; Southern Alberta Fly Fishing

Outfitters, http://albertaflyfish.com, which N2H2

blocked as “Pornography”; and “Nature and Culture

Conscious Travel,” a tour operator in Namibia, http://

www.trans-namibia-tours.com, which was categorized

as “Pornography” by N2H2.

The filtering programs also miscategorized a

large number of sports Web sites. These included:

a site devoted to Willie O’Ree, the first African Ameri-

can player in the National Hockey League, http

www. missioncreep.com/mw/oree.html, which Websense

blocked under its “Nudity” category; the home page of

the Sydney University Australian Football Club, http://

www.tek.com.au/suafc, which N2H2 blocked as “Adults

Only, Pornography,” Smartfilter blocked as “Sex,”

Cyber Patrol blocked as “Adult/Sexually Explicit” and

Websense blocked as “Sex”; and a fan’s page devoted to

the Toronto Maple Leafs hockey team, http://Awww.

torontomapleleafs.atmypage.com, which N2H2 blocked

under the “Pornography” category.

90a

7. Conclusion: The Effectiveness of Filtering

iety of means of

ic libraries have adopted a variety of mes

1 with problems created by the a. “4

Internet access. The large amount of sexually explici

speech that is freely available on the Internet has, >

varying degrees, led to patron complaints about —

matters as unsought exposure to offensive —

incidents of staff and patron harassment by individ -

viewing sexually explicit content on the Internet, —

the use of library computers to access illegal —

such as child pornography. In some libraries, you —

library patrons have persistently attempted to use the

Internet to access hardcore pornography.

ublic libraries that have responded to these

41 using software filters have — such

filters to provide a relatively effective means 0 =

venting patrons from accessing sexually explici

material on the Internet. Nonetheless, out of the *

universe of speech on the Internet falling within —

filtering products’ category definitions, the —

incorrectly fail to block a substantial amount ol —

Thus, software filters have not completely —

the problems that publie libraries have sought ;

address by using the filters, as evidenced by frequen

instances of underblocking. Nor is there any —

tive evidence of the relative effectiveness of filters an

the alternatives to filters that are also —

prevent patrons from accessing illegal content on the

Internet.

i i Ithough

more importantly (for this case), althou

BD filters provide a relatively cheap and effective,

albeit imperfect, means for public libraries to prevent

patrons from accessing speech that falls within the

9la

filters’ category definitions, we find that commercially

available filtering programs erroneously block a huge

amount of speech that is protected by the First

Amendment. Any currently available filtering product

that is reasonably effective in preventing users from

accessing content within the filter’s category definitions

will necessarily block countless thousands of Web

pages, the content of which does not match the filtering

company’s category definitions, much less the legal

definitions of obscenity, child pornography, or harmful

to minors. Even Finnell, an expert witness for the

defendants, found that between 6% and 15% of the

blocked Web sites in the public libraries that he

analyzed did not contain content that meets even the

filtering products’ own definitions of sexually explicit

content, let alone CIPA’s definitions.

This phenomenon occurs for a number of reasons

explicated in the more detailed findings of fact supra.

These include limitations on filtering companies’ ability

to: (1) harvest Web

This text is long and has been trimmed here. Open the source document for the complete record.