On Apr 15, 2008, at 2:26 PM, marius schebella wrote:

zmoelnig@iem.at wrote:

...

Quoting "SourceForge.net" noreply@sourceforge.net:

...

Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: pdpedia

...

Status: Closed Resolution: Invalid

Submitted By: oli44 (oli44) Assigned to: Nobody/Anonymous (nobody) Summary: pdpedia Captcha

Any admin can easily delete those junk posts with the 'delete' tab and also protect problem pages. A captcha might help, I haven't seen one for mediawiki before.

a captcha will help in about 97% probably even more (this is just a wild guess)

...

Check the "recent changes" log now, and you'll see the spam is gone. :)

hmm, your approach seems to deal with the problem but does nothing to prevent it. what will happen when the next attack starts? will we just have to wait for "any admin" to "easily delete those junk posts" again?

as far as i recall, when pdpedia was installed there was some consensus about using captchas when a spamming problem would appear. are there any technical or organisatorial problems in using them? (it seems like you have some reasons against using captchas, but you haven't told us yet)

I think we even agreed that in a case when sites get spammed, that only registered users were allowed to edit pages. I still would recommend that, I think then we don't need captchas. marius.

Spambots can easily register themselves and post, it happens a lot. Then you have to deal with managing spam users. When spambots can post anonymously, then it is easy to manage the removal of it. Managing spam users would be a lot more work.

.hc

------------------------------------------------------------------------ ----

Man has survived hitherto because he was too ignorant to know how to realize his wishes. Now that he can realize them, he must either change them, or perish. -William Carlos Williams

Hans-Christoph Steiner wrote:

You cannot stop spam ever. Period. Captchas are annoying and are far from 97% effective. I have captchas on my blog and I get a lot of automated user registrations anyway.

interesting. for me spamming has stopped completely after i introduced captchas on the sites i administrate. do you have any ideas why captchas do not work for you? - are the captchas readable by machines - are human spammers deciphering the captchas - are humans deciphering captchas for spammers (unknowingly) - do spammers exploit a vulnerability in your blog, that allows them to bypass the captchas?

btw, i had a look at your blog and i don't see any captchas.

a log-in does not prevent bots, even if you require a valid email-address to activate an account.

spammers registered about 3000 user-accounts at pure-data.info and posted junk, before i installed captchas. since then i cannot remember any spam on that site.

I think wikipedia people are very smart about this issue. The real focus should not be to prevent spamming at all costs. If you do that, you end up working more, and you lose a lot of contributions.

obviously this is something we want to prevent.

The real focus of this effort should be how to manage the problem in a way that takes the least work and provides the most benefit. If we

again, from my personal experience, it took me no more than 6 hours to install the captchas system (once i had found a reasonable one). chances are high that it took me about 2 hours (but this has been a while ago, i cannot remember it accurately). i haven't spent any minute since then on the system.

leaving the work for pdpedia-admins will eventually result in more time spent on dealing with the problem.

(ah i realize the error in my little calculation: if you add up all the time users have to spend deciphering captchas, the effort spent in reducing spam might indeed be higher when using captchas)

have a number of admins active on pdpedia, then it is really quite easy to use the 'delete' feature to get rid of spam. For pages that get spammed a lot, there is the 'protect' feature.

there are 2 problems i see here: - "if we have a number of admins active"; indeed, _if_ we have them. it took quite a while for one of the admins to fix the content. for whatever reasons, this admin has been _you_ - using the "protect" feature is exactly the opposite of a wikipage as it forbids editing. how does this relate to "working more, and you lose a lot of contributions" (ah right: the difference is, that you lose _all_ contributions)

finally, an example from the real-world:

looking at the "recent changes" of wikipedia right now, i don't see any junk. looking at the "recent changes" on pdpedia right now, the last 6 entries are junk content for "Template:jmmmmp_author" (this happended _after_ you did the fixes) according to the logs you have had "protected" this object. either protection is not very efficient or somebody has removed protection (i guess the latter is the case, as you seem to have deleted this object after you activated protection)

fgmasdr IOhannes