sciencegeek: (V for Vendetta II)
sciencegeek ([personal profile] sciencegeek) wrote2007-12-10 02:33 pm
Entry tags:

Whoa. What? Wait. What?

I like to hope that LJ has their hearts in the right place, I do...but seriously?

Banned Interest Terms

Excerpt

I'm told this is done by "substring search"--so certain root words are blocked, and every derivative that includes them. There are two categories: blocked words, and blocked combinations. If you search for one of the blocked terms, you get a message that says:

Error
Sorry, we're unable to help you find users matching the interests you've provided.
instead of the the normal list of comms & journals, or note that nobody has this interest, but you could start a comm with it or add it to your journal.

The combinations seem to attempt to address child/incest sexuality issues; the others seem selected to address racism and certain WWII issues.

The intent: Block the objectionable terms, and new interests people make that attach the objectionable terms to a new word.
The result: Hundreds of innocuous interests are blocked from search.

Among the blocked terms are:
crackerjacks,doo-wop music, hospices, baby grapes, Rollercoaster Tycoon, wharfage,suspicions, unisex baby clothes, beaneries, despicable liars,pedagogics, pachinko games, spice girls, gobbledegook.

Intentionally blocked are:
childrape, sex with little girls, kiddie porn, niggers, kikes, god hatesfags, gook killers, ragheads, fucking your mother, heil hitler, stupidgooks, kkk.


more here

I think they have the best intentions, I do...but like much of LJs recent shananingans, it fails in execution.
florahart: (Default)

[personal profile] florahart 2007-12-10 07:51 pm (UTC)(link)
The thing is, this is a predictable problem. Anyone who knows ANYthing about how search strings work, knows that a string that is short and nonexact (that is, any time in a search engine you might use, depending on which one, a * or ? or other truncation character in order to indicate "this word or any others that contain it" you are hugely likely to get a ton of other words you don't mean. I mean, like, libr* will get you Libra and libretto, as well as librarian, librarians, library, librarianship, (and if it's *libr* it'll also get calibrate, etc) so if you are specifically searching for library stuff, you will have to make your search string librar*

*I* know this, and I am not a coder. At a guess, they chose to implement quick and dirty code and hope no one would notice, or hope no one would notice until they had time to do a better job, because there is NO WAY they should have found this remotely difficult to predict--and if they did, they have bigger database-management problems than anything they've managed yet. I mean. I can see why they would make this choice (to prevent alarming right-wing groups from coming in, searching for chan/buggery/other scary things and then flagging everyone they find again), but HONESTLY, there was no reason to hide it. Now, it looks shady. Better: to say, hey, we've tossed a filter into the search code in order to make it harder for vigilantes to go finding you all by interests they find creepy, but it has a bad case of imperfect because in order to implement it quickly, we had to make it wipe out wide swaths of searchability. Code Monkey Joe (username joethemonkey) is working on writing better code but meanwhile we're going to live with this.

Which, then of course they would have had people asking, okay, it's been 3 months, what the hell, but at least it wouldn't look so flipping conspiratorial and/or idiotic.

[identity profile] sciencegeek.livejournal.com 2007-12-10 07:57 pm (UTC)(link)
NO WAY they should have found this remotely difficult to predict--and if they did, they have bigger database-management problems than anything they've managed yet.

heh, agreed. And I see why they did it too. And I see why they might want some of those terms unsearchable, but the timing is shit, and as you say, if they didn't think that predict this exact problem then...they have huge problems.

[identity profile] skuf.livejournal.com 2007-12-11 07:06 am (UTC)(link)
See, my main beef is that I don't think any terms should be made unsearchable! Like, I understand "genocide" is intentionally blocked - wtf, how many people listing or searching for "genocide" are going to be PRO-genocide? It's ridiculous!

[identity profile] sciencegeek.livejournal.com 2007-12-11 12:51 pm (UTC)(link)
I don't think what they did was right, but I sort of see why they did it (and, with the stupid antics this summer, I am unsurprised).

Like, I understand "genocide" is intentionally blocked - wtf, how many people listing or searching for "genocide" are going to be PRO-genocide? It's ridiculous!
I know! And it's a derivative thing as well - spic is blocked therefore you can't look up things like "spicy foods" or "spice girls". Which is just about the stupidest thing in, oh ever.