FC: Will Doherty replies to Bruce Taylor on library filtering

[Declan McCullagh <declan () well com>]

Previous Politech message:
http://www.politechbot.com/p-04533.html

---

Date: Mon, 10 Mar 2003 15:32:29 -0800
To: declan () well com
From: Will Doherty <doherty () onlinepolicy org>
Subject: Re: FC: Bruce Taylor on library filtering case: Justice Dept
  won?
Cc: doherty () onlinepolicy org
In-Reply-To: <5.1.1.6.0.20030310173126.0172e808 () mail well com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"; format=flowed

Dear Declan and Bruce (and politech readers if this is forwarded),

Keyword-based filters often block search engine results.

Even if search result listings are not blocked, there is
often not enough information available in the brief
search result listing to determine whether or not
the search result should have been blocked and many
search results are in fact inappropriately blocked
for many different logical and entirely unpredictable
reasons.

At least one filtering vendor blocked its own website
front page.

You can see examples of such behavior in the Online Policy Group's
"Online Oddities and Atrocities Museum" under construction at
http://www.onlinepolicy.org/research/museum.shtml (your
submissions to the museum are welcome).

Internet filtering products simply don't and can't work
since they aren't capable of interpreting human language
in context.

School children in the United States are going to have
Swiss cheese for brains due to the randomly inaccessible
information they will miss from their education due to
Internet blocking. This will be true even if the U.S.
Supreme Court upholds CIPA for libraries, since schools
weren't covered in this legal challenge.

One Internet with Equal Access for All,

Will Doherty
Online Policy Group, Inc.
http://www.onlinepolicy.org
-------

---

To: declan () well com
Cc: politech () politechbot com, BruceTaylor () NationalLawCenter org
Date: Tue, 11 Mar 2003 02:24:44 -0500
Subject: Re: FC: Bruce Taylor on library filtering case: Justice Dept won?
From: terry.s () juno com

> out which ones the filter will block.  There are no "secret" blocks
> by filters.  A filter always tells you when you can't see a site, but
> you have to ask to see the site before the filter is asked if it's
blocked.

To the extent there's bias in blacklists, and many vendors treat those
lists as proprietary secrets, librarians cannot be aware of the patterns
in censored content, while citizens cannot inquire and discover same to
review for possible biases, short of obtaining and hacking encryption in
those elements of such software, or doing massive amounts of testing for
blocked URLs.  Every blacklist used in a tax funded system should be FOIA
available.  That includes whitelists, since creative bigots have been
known to claim they don't deny blacks access to busses or lunch counters,
but only allow whites.


> Analogize to library card catalogues:
>
> A library has a white card for every book they have.
> The card catalogue doesn't have a card for every book in print on a
> topic and doesn't have a card for the books they don't have.
> When users go to the card catalogue, we only learn which books the
> library has on a topic, not all the other books ever printed on that
> topic.
> If they had a white card for the books they have and a red card for
> the ones they don't, at least patrons would know what else exists on
> a topic.

In past history, this might have been a valid analogy.  Due to shared
databases, the move from paper card catalogs to searchable electronic
ones, interlibrary loans, and access to additional databases such as
"books in print" from library computers, modern libraries do have in
effect three sets of cards, actually more inclusive than what Bruce's
model suggests.  They have those "white cards", "red cards" for what's
available from other libraries on loan, constituting both in print and
out of print materials not always easy to purchase if one so chose, and
"yellow cards" for other books which the instant or some other library
might or might not purchase if requested, or which might be otherwise
available to the patron to buy or borrow from other sources.

That changed reality of modern library operations makes Bruce's model
flawed in other ways.  It costs staff labor and transportation to operate
interlibrary loan systems.  Presumably were shelf space and acquisition
costs not issues, most general use libraries would have 100% of items on
the "red" list, additional copies of popular works on the "white" list,
and many if not most items from the "yellow" list as well.  Web sites
equivalent to those lists have zero incremental budget or space
requirement to carry multiple copies of popular works and all other
patron desired materials.  In effect that means the net enables expanding
the "white" list, adding the "red" list at lower cost than interlibrary
loans (reversing traditional cost models), and simply making every item
on the "yellow" list available for less cost than staff time alone that
would be required to review patron requests for collection additions in
print form.

The only forms of libraries which traditionally try to avoid having
expansive collections bounded only by cost and space are topical focus
libraries.  Those might be found in a university science or arts
department, a court house, or an elementary school where reader levels
10-14+ would be wasted while complicating the tasks of users finding
useful materials easily.

These changed cost models offer additional opportunities to remedy past
issues of discriminatory practices, which in turn result in absurdities
under Bruce's model.  Most US libraries carry almost exclusively English
language works.  Some carry Spanish or other language books based on
significant local community constituencies.  Most library patrons with
interests in minimally represented languages are left to fend for
themselves.  This economic model of not designating a legal national
language (with all the religious and nationality biases that would carry)
while catering to more prolific minorities treats members of smaller
minorities as if less than citizens.  The virtually infinite expansion of
shelf capacity without acquisition costs allows carrying every available
foreign language version of every book in the catalog or loan or in print
systems, plus allows near instant translations (albeit with defects) to
nearly any other language.  No longer would people whose first languages
were Asian, African tribal, Portuguese, French, German, Hebrew, Arabic,
etc. expect to find versions in their primary languages absent.

This raises an interesting issue as to most porn sites (whatever that is,
with tribute to Potter Stewart's respect for the inability of US law to
promulgate a neutral and concrete definition).  Were we to tolerate
English and possibly other prevalent language based filters to impose
Abrahamic religious prejudices backed by defective past Supreme Court
rulings demeaning sexuality, would English speaking citizens be denied
access to types of sites which persons working in less common foreign
languages might find not blacklisted?

What biases would be present in a Jain friendly filter, as a religious
perspective which goes beyond values of most pacifist Earth friendly
sexuality-positive Neopagans, but with an ascetic side more aligned with
Calvinist religions?  Depiction of violence would clearly be offensive,
whether advocacy of war, the artifacts of "War on Select Drug Rights"
black markets, or the killing of freshly plucked vegetables, never mind
animal species of any kind.  Raid insecticide ads would be gone.  Florida
oranges or milk might pass, while Gutrot Arches or brocoli would be out.
Cars and plastic goods would be gone, as would ads for clothes if the
ideals of digambaras Jains were respected.  As such, images of humans
wearing clothes might be seen as offensive, though only the mendicants
(clergy) of that sect actually vow to not own clothes.  Sexuality might
be evaluated as to its healthy in balance and sometimes spiritual forms,
and depicted openly.  However, if we look to tantra or common Neopagan
methods to determine what's healthy sexuality or not, that's an
individual balance issue, where what's out of line is any third party
imposing his judgement in place of the individual's guidance to his own
path.

That brings us full circle, to recognize that there are no neutral
criteria for content based censorship.  Every possible blacklist is based
on popular prejudices inclusive of religion, ethnicity, artistic ideas,
or political viewpoint, and so fails the basic tests of Cohen v CA or
Hess v IN.  Had our courts been entirely honest, Miller v CA would have
collapsed as of 1868 (14th extending 1st inside states), and Pacifica
indecency would have been treated as anyone's right to interpret for
himself outside the US legal system, just as Justice Stewart suggested
for "pornography", not established in precedent in the 1970's.

The only rational conclusion is that whether out of economic or political
power greed, or prejudices of zealotry, the core intent of every possible
censorship effort, much like Secondary Effects Doctrine, is to find
facially legal excuses for blatantly illegal discrimination.  It costs a
library less for a patron to learn about "making sparks" with dripping
wet pink pussy online, than it costs to stock a spare Boy Scout Manual
that teaches about what for many are less exciting sparks.



Terry



"Censorship is the most insidious form of hate speech."




-------------------------------------------------------------------------
POLITECH -- Declan McCullagh's politics and technology mailing list
You may redistribute this message freely if you include this notice.
To subscribe to Politech: http://www.politechbot.com/info/subscribe.html
This message is archived at http://www.politechbot.com/
Like Politech? Make a donation here: http://www.politechbot.com/donate/
-------------------------------------------------------------------------
Declan McCullagh's photographs are at http://www.mccullagh.org/
-------------------------------------------------------------------------