June 25, 2003

"CIPA-compliant" library censorware

The idea of minimal, "open-source", library-specific censorware is being widely discussed (see, e.g. Edward Felten's comments - thanks Donna)

Here's the problem:

1) If any library wanted to play challenge-the-law, all they would need to do is sit back and say "Give us the specific, judicially-decided, URLs to be banned, and we'll ban them -- but not one URL more." And then wait for the compliance lawsuit to be brought. Very simple.

2) If they don't want to be challenging the law, why would they undertake what will certainly be a major PR hassle? That is, anyone can come up with harsh-but-not-illegal sites and say "Library X allows these PORNOGRAPHY sites to be viewed!". So do they get added to the blacklist or not? You mean the library is going to stand up to a constant barrage of bad PR like this? If they were willing to do that, we'd be in case #1.

Two words: Robert Mapplethorpe.
Blacklisted or not? Think through your answer in either case.

What happens when the "North American Man-Boy Love Association" asks to be whitelisted?

The idea of Open-Source Censorware (more accurately, an Open-Source Censorware Blacklist) is one which is very appealing from 10,000 feet. But it falls apart on any close examination.

OpenCensorware is far more work than may be apparent.

Here's the most well-known people who are trying it:


Heard of them? No? Consider there's reasons why.

By the way, the Australians tried this idea too:


"Announcing the GnU Internet Lust Terminator, an open-source censorware proxy that only filters ABA-supplied banned URLs.

The software is being developed by Zem for 2600 Australia and will be eventually submitted to the IIA for inclusion as an approved filter"

The Australian government didn't approve it.

[Update 6/26 - I've also suggested privoxy (http://www.privoxy.org/) ]

Before people write back, here's my challenge:

Don't tell me this is such a great idea. Find libraries who will use it who agree it's such a great idea!

[Disclaimer - I said to one proponent of this idea that I'd help make it happen, if he could find libraries which wanted it, and funding for it, as part of a challenge above]

By Seth Finkelstein | posted in censorware , infothought | on June 25, 2003 11:25 AM (Infothought permalink) | Followups
Seth Finkelstein's Infothought blog (Wikipedia, Google, censorware, and an inside view of net-politics) - Syndicate site (subscribe, RSS)

Subscribe with Bloglines      Subscribe in NewsGator Online  Google Reader or Homepage


If I read this right, your problem is who builds the black list, and what criteria is used to determine its contents. What if it was built by the users themselves?

I'm fuzzy on the details (yes, I know where the devil lives), but what if the 'blacklist' was a sort of peer->peer ratings system? Then the libraries would control the list amongst themselves.

Sort of like some of the interesting user-based spam filtering proposals I've seen around...

You could even let local users have overrides to tailor it to their local needs.

Of course it's more work for the librarians..

Posted by: Tom Haviland at June 25, 2003 12:05 PM

Oh boy - calling all Christian Coalition members!

Spam, as the saying goes, is about consent, not
content. Whereas this is the essence of a
content-based determination.

Posted by: Seth Finkelstein at June 25, 2003 12:13 PM

Could be. Only it wouldn't be all "members" it would be all "librarians"

Posted by: Tom Haviland at June 25, 2003 04:34 PM

Felten's proposal was designed to fix two problems: censorware that blocks political (but not indecent/obscene) speech; and censorware that does not identify what sites it blocks. His idea is perfectly consistent with censorware that is very cautious and conservative about indecent material.

I don't see any reason why an open-source censorware program would have more problems with indecent material than a closed-source program. Every one of these programs has holes in them, and the solution is to get frequent updates.

Just exposing the list of sites would be a huge step forward. As far as policies, let a thousand flowers bloom - i.e., let there be a variety of policies for librarians to choose from, that are more or less strict. This way they can satisfy their local community standards regarding indecent material without the objectionable side effects mentioned by Felten.

Posted by: John at June 26, 2003 02:41 AM


Don't tell me this is such a great idea. Find libraries who will use it who agree it's such a great idea!

Posted by: Seth Finkelstein at June 26, 2003 02:44 AM


I think that such librarians may just exist.

I posted a similar idea at Web4Lib--and I swear, I didn't read yours or Felten's comment first--and got an almost immediate response from Cindy Murdock at Meadville Public Library/Crawford County Federated Library System

Here's the reference:

Posted by: walt crawford at June 26, 2003 12:36 PM

The concern over open/closed lists seems to me like arguing over the technical details of how to best implement a much larger problem. First, if the lists are open, who really cares. With open lists we can do data mining and generate controversy over who supports what politics. But who blocks what only matters under two conditions. First, it matters when the filter cannot be circumvented by legitimate means. As long as we can turn it off we can get to the information we want when we want it. So the “censorship” of the filter becomes a non-issue. The problem is when we cannot turn it off. That is when the filter becomes censorship. It does not matter whether the list is open or closed. If the filter cannot be circumvented, then all lists count as censorship. If the filter can be circumvented, then no list can really count as censorship.

Second, and more important, the filter's blacklist matters when the user is uninformed when the filter blocks something. Does the filter give a banner page that says “This has been blocked. Ask [your parents] for assistance.” Or does the filter give a “page not found” error or silently let the request time out. If the user knows the filter is at work, and the filter can be circumvented, all is well. If the filter cannot be circumvented, she at least knows to try again somewhere else. But if the user is uninformed that the filter is at work, that is when the filter becomes real censorship. Once again, it does not matter whether the list is open or closed. If the user is kept uninformed of the filter's operation, then all lists count as censorship. If the user is always informed, and the filter can be circumvented as needed, then no list can really count as censorship.

So let's not waste our energy on open vs. closed lists. The list does not really matter. We should make sure the filter can be circumvented and users know when the filter blocks something when it happens. In fact, I could argue that an overly conservative list actually contributes less to censorship. If requests to circumvent the filter are reasonably common and obviously reasonable, people will feel less intimidated about needing to do so.

Posted by: Paul Waak at November 4, 2003 07:18 PM