This page was written 72 hours after the body of this report was made public. In that time, Microsystems/The Learning Company has, as predicted, conducted an extensive damage-control effort. First, they published a press release. Please examine it in light of the information previously presented.
Microsystems denies none of the blocks we listed. And they attempt to spin the issue to be one of children accessing porn. The word "children" is used seven times in a short release. Yet Microsystems markets the same list in a nearly-identical program for use on adults in their corporate and proxy server versions, so why all this emphasis on children? Because it plays better with the press.
Microsystems says in their Dec. 23 release:
While the software is used by some libraries based on specific sentiment in their communities, The Learning Company does not market the Cyber Patrol software to libraries.
But of course they do. See their own February 1997 press release:
Microsystems Software demonstrates Cyber Patrol for Boston Public Library . . . "Cyber Patrol allows the City of Boston to protect children from inappropriate content in cyberspace while safeguarding our fundamental freedom of speech," said Nigel Spicer, President of Microsystems Software Inc., who will demonstrate the software on Thursday morning. . . . Librarians, teachers and parents have the ability to add sites they want blocked or delete sites they want children to be able to access. The filter can be turned on or off, loaded on some computers and not others and used to implement library policy.
So what exactly does "does not market [Cyber Patrol] to libraries" mean? That they don't go door to door drumming up sales? It certainly doesn't mean that they will refuse to sell to libraries, unless Microsystems and/or The Learning Company (which owns Microsystems) have made an abrupt about-face in the past few days. Apparently having the company President show up and provide a product demonstration is not "marketing" it, it's, uh, something else.
So, somewhat disappointingly, we cannot have much of a dialog with Microsystems about the mandatory use of this software on adults by corporations, libraries, and governments. It's a shame that all Microsystems can talk about is children, because that's the only area where they feel they are on secure moral ground, where they feel they can deflect criticism.
More importantly, we should look at the list of sites blocked under Cyber Patrol's FullNude and/or SexActs categories. Besides the press release, over the last 72 hours, they have put out two new rebuilds of their blacklist, and updated their "HotNot" list several times as well. We covered a total of 67 specific block entries. Each of these is a partial URL that blocks all URLs matching it, and each of these blocks covered at least a full domain's worth of information, with some of the blocks banning thousands upon thousands of webpages. The single block on members.tripod.com, for example, blocks over 700,000 individual user websites.
Using the latest edition of the blacklist downloaded from Microsystems' servers, we've reviewed the sites. Microsystems has had time to correct their errors, and we were curious about the results.
Twelve blocks remained of the 67 blocked URLs/domains, an acknowledgement by Microsystems that we were correct in pointing out 55 of their mistakes. But although we made a few mistakes of our own, Microsystems' refusal to unblock many of the remaining twelve further evidences their carelessness or their willfulness, since many of the twelve clearly should be unblocked.
The eight blocks which very clearly should not be blocked are:
Not only did Microsystems fail to unblock this, but their presumed inspection of the site was too cursory to click on "Hot Links" on the homepage and find the adult sites that should have been added to their CyberNOT list. Double whammy.
The four remaining blocks include two which contain a small amount of explicit material (but are still overbroad blocks) and two which are indeed legitimate blocks and which we should not have listed in the first place:
And there's a commercial nudie-pics outfit. What we didn't catch on the first run-through is that that site and the mall are both run by the same guy. Microsystems didn't go that in-depth - they saw the sexy-pics site and banned everything. The one page with sexual content is http://classact.com/public/screen.htm, and it has zero nudity available without paying a fee and nearly-zero sexual text.
It also has some very bland sexually-suggestive text on it. Basically they are selling Club Med-style vacations, "wine, women and song." A search for nude pictures turned up one racy animated .gif, which seems to meet the FullNude definition, but Cyber Patrol could have blocked this page individually rather than scooping in the entire extensive site. Microsystems made no changes to their block of the entire site.
These two sites are commercial picture sites. Both of them reasonably meet the Cyber Patrol criteria, and shouldn't have been listed as wrongful blocks originally.
If anything, this exemplifies the problems with scanning sites quickly to give them a thumbs-up or down. We were preparing these pages for publication and intense examination and trying quite hard within the time available not to make any errors. And we weren't successful. When we end up with errors examining a short list of sites for open publication, we wonder how any accuracy can be achieved examining the World Wide Web for a secret database where your errors generally resemble Jimmy Hoffa - out of sight and out of mind.
So the bottom line, for those who try to obfuscate the real issue by playing silly numbers games, is The Censorware Project 64, Microsystems 2, and that's without yet looking at all the improperly blocked newsgroups.
So why don't we review some of the banned groups:
The entire rec.games.* hierarchy, 100 groups concerned with games of all types, continues to be banned as containing illegal, violent, profane, and intolerant content. Misc.headlines and the misc.health.* hierarchy (the block is on "misc.hea") continue to be banned under Violence/Profanity and SexActs. The news.groups.* tree is still banned, ensuring that new users who want to find out what newsgroups cover a specific topic will find it challenging to do so. Rec.hunting.* is still banned under Militant/Extreme. The soc.* hierarchy is still hard hit - groups for soc.support.*, for example, are still banned as profane and intolerant, and soc.feminism is still banned for Violence/Profanity and SexActs, despite the fact that it's a moderated group. Alt.coven continues to be banned, despite Microsystem's promise earlier this year to quit discriminating against non-mainstream religions. Alt.censorship is still censored, in a fine display of irony. The alt.cyber* tree continues to be banned. Feminism? Banned. Journalism? Banned. The 220 mostly-unrelated groups that match the pattern alt.sup*? All banned.
In fact, Microsystems has not made any visible changes whatsoever to their newsgroup blocks. Not one. We put out several hundred newsgroups (the product of a hour-long check using Cyber Patrol) that were assuredly wrongly blocked or at least deserved another look, and Microsystems didn't bother to review any of them.
In sum: Microsystems reviewed some of the bad blocks we presented for the web, although they failed to unblock many that they should have. They did not undertake any review of the many newsgroups wrongly Banned By Cyber Patrol.
In our opinion, these results should give anyone pause. Why those websites are banned cannot be explained by Microsystems, despite their claims that all sites are human-reviewed before being added to the blacklist. Their review procedures are cursory, limited to damage-control efforts only, and any of these sites could, of course, be added back to their list at any time. They seem to have made a command decision to continue banning members.tripod.com, since that is easier than trying to evaluate 1.4 million constantly-changing web pages hosted by that provider.
If Cyber Patrol can't deal with 67 specific blocks handed to them on a platter and with public scrutiny, how well can they maintain a database with (currently) fifty-eight thousand secret, unreviewed entries?
back to top