On 23/11/05, Ilya N. ilyanep@gmail.com wrote:
Consider a bot (or many humans) going into the validation system and sending totally bogus (or even offensive) entries into it.
My understanding of the validation feature was that it was a "0,1,2,3,4,5 - pick one" type system, rather than a "Please leave comments" one, meaning that whilst you could game it with a bot - and no doubt that'll be attempted by a rather cunning linkspammer or three - you couldn't fill it with offensive entries. (Or will there be a comments field? My connection seems to be a bit unfriendly with meta. this morning, and I haven't been checking lately...)
But that's what the testing phase is for. We _want_ people to try and spam it, try to vandalise it, do all this stuff. Then we can figure out what's good in the validation system and what's bad - perhaps we could only make comments visible to logged-in users, or to admins, if there's a problem with junk filling them up, or implement a filter to only show "useful" comments.
(How do we define "useful"? I don't know. Maybe we have a problem with people spamming links in comments fields, so only make anything without a URL publicly visible. Maybe we have a problem with people just writing "fucking crap", so we put in a seven-words filter. We don't know yet, but it doesn't seem beyond the wit of man to solve what results...)
-- - Andrew Gray andrew.gray@dunelm.org.uk