Restoring the original content is a bit tricky with the ALU CLiki, which does not currently provide links to previous page versions. I seem to remember that such links used to be available.
If enough Lispers from all over the world and time zones check the sites, we may be able to reduce the effectiveness of spam, or at least raise the cost of producing it.
Paolo Amoroso <amor...@mclink.it> wrote: > Restoring the original content is a bit tricky with the ALU CLiki, > which does not currently provide links to previous page versions. I > seem to remember that such links used to be available.
> If enough Lispers from all over the world and time zones check the > sites, we may be able to reduce the effectiveness of spam, or at least > raise the cost of producing it.
Lisp folks with WikiClues are also, of course, invited to contribute to the growing Wikipedia articles on the subject, including:
(Or, of course, the non-English Wikipedias' articles on the same subject....)
-- Karl A. Krueger <kkrue...@example.edu> { s/example/whoi/ }
Every program has at least one bug and can be shortened by at least one line. By induction, every program can be reduced to one line which does not work.
On 2004-12-17 17:18:24 +0100, Paolo Amoroso <amor...@mclink.it> said:
> Restoring the original content is a bit tricky with the ALU CLiki, > which does not currently provide links to previous page versions. I > seem to remember that such links used to be available.
They are not available but you can append ?v=x , where x the the version, to the url.
Engelke Eschner <enge...@tekai.org> writes: > On 2004-12-17 17:18:24 +0100, Paolo Amoroso <amor...@mclink.it> said:
>> Restoring the original content is a bit tricky with the ALU CLiki, >> which does not currently provide links to previous page versions. I >> seem to remember that such links used to be available.
> They are not available but you can append ?v=x , where x the the
> Restoring the original content is a bit tricky with the ALU CLiki, > which does not currently provide links to previous page versions. I > seem to remember that such links used to be available.
> If enough Lispers from all over the world and time zones check the > sites, we may be able to reduce the effectiveness of spam, or at least > raise the cost of producing it.
I've no idea whether that's a spammer-detection procedure that's likely to work usefully in the future, but it doesn't seem unlikely. (If it works for a few months, then probably it's worthwhile.)
Paolo Amoroso wrote: > From time to time, CLiki site are flooded by spam. This is currently > the case for the ALU CLiki, as you can see from the change log:
The obvious solution that reduces spam consists of measures that are implemented on various online systems:
Before creating content, users are ``authenticated'' to an e-mail address by creating a password-protected account that is activated by responding to an e-mail and responding to an image-understanding challenge.
Whenever submitting new public content, users must be logged in to an activated account, and additionally respond to an image-understanding challenge.
> Whenever submitting new public content, users must be logged in to an > activated account, and additionally respond to an image-understanding > challenge.
>> On 2004-12-17 17:18:24 +0100, Paolo Amoroso <amor...@mclink.it> said:
>>> Restoring the original content is a bit tricky with the ALU CLiki, >>> which does not currently provide links to previous page versions. I >>> seem to remember that such links used to be available.
>> They are not available but you can append ?v=x , where x the the
Stefan Scholl <ste...@no-spoon.de> writes: > On 2004-12-18 02:08:42, Kaz Kylheku wrote:
> > Whenever submitting new public content, users must be logged in to an > > activated account, and additionally respond to an image-understanding > > challenge.
> That wouldn't be a wiki anymore.
The image-understanding challenge by itself may be enough to eliminate automated spam. Although I don't think that is friendly to blind people or people who have images turned off.
How about if you just ask, "Are you a human?" If the answer is "yes", they can update the page. If the answer is "no", they can't.
-- An ideal world is left as an excercise to the reader. --- Paul Graham, On Lisp 8.1
David Steuber wrote: > Stefan Scholl <ste...@no-spoon.de> writes: > > On 2004-12-18 02:08:42, Kaz Kylheku wrote: > > > Whenever submitting new public content, users must be > > > logged in to an activated account, and additionally > > > respond to an image-understanding challenge.
> > That wouldn't be a wiki anymore.
> The image-understanding challenge by itself may be enough to > eliminate automated spam. Although I don't think that is > friendly to blind people or people who have images turned off.
> How about if you just ask, "Are you a human?" If the answer is > "yes", they can update the page. If the answer is "no", they can't.
Just so you guys don't waste your time on this in vain, the ALU wiki problem has nothing (nada, zero) to do with technical issues. Paolo probably knows this, and suggested the only obvious solution.
I personally do not think Lisp users should feel at all responsible to contribute labor to this wiki unless they are honestly told the specifics of why this spam is an issue. I'm not dissing Paolo, I think he's just being polite when faced with a totally bullshit situation. MfG, Tayssir
Karl A. Krueger wrote: > Paolo Amoroso <amor...@mclink.it> wrote: > > Restoring the original content is a bit tricky with the ALU CLiki, > > which does not currently provide links to previous page versions. > > I seem to remember that such links used to be available.
> > If enough Lispers from all over the world and time zones check the > > sites, we may be able to reduce the effectiveness of spam, or at > > least raise the cost of producing it.
> Lisp folks with WikiClues are also, of course, invited to contribute > to the growing Wikipedia articles on the subject, including:
I've always been unhappy with descriptions of sexps; they seem to get lost in technical descriptions of cons cells, and obscure the point. So I wrote much of this: http://c2.com/cgi/wiki?EssExpressions Perhaps I can tidy something like it up for Wikipedia.
> Karl A. Krueger wrote: >> Paolo Amoroso <amor...@mclink.it> wrote: >> > Restoring the original content is a bit tricky with the ALU CLiki, >> > which does not currently provide links to previous page versions. >> > I seem to remember that such links used to be available.
>> > If enough Lispers from all over the world and time zones check the >> > sites, we may be able to reduce the effectiveness of spam, or at >> > least raise the cost of producing it.
>> Lisp folks with WikiClues are also, of course, invited to contribute >> to the growing Wikipedia articles on the subject, including:
> I've always been unhappy with descriptions of sexps; they seem to get > lost in technical descriptions of cons cells, and obscure the point. So > I wrote much of this: > http://c2.com/cgi/wiki?EssExpressions > Perhaps I can tidy something like it up for Wikipedia.
Do you have a reference for wherever Rivest proposed replacing cons cells with what sounds like tuples?
Peter Seibel wrote: > "Tayssir John Gabbour" <tayss_te...@yahoo.com> writes: >> I've always been unhappy with descriptions of sexps; they seem to >> get lost in technical descriptions of cons cells, and obscure the >> point. So I wrote much of this: >> http://c2.com/cgi/wiki?EssExpressions >> Perhaps I can tidy something like it up for Wikipedia.
> Do you have a reference for wherever Rivest proposed replacing cons > cells with what sounds like tuples?
I didn't write that part, and haven't the foggiest. Maybe someone can ask about it there.
Incidentally, I recall the guy who wrote that was one of the heavy flamers, so I suspect he was trying to paint lisp as some sort of priesthood with sacrileges. But it's a helluva lot better than what went on before...
Tayssir John Gabbour wrote: > Just so you guys don't waste your time on this in vain, the ALU wiki > problem has nothing (nada, zero) to do with technical issues. Paolo > probably knows this, and suggested the only obvious solution. > I personally do not think Lisp users should feel at all responsible to > contribute labor to this wiki unless they are honestly told the > specifics of why this spam is an issue. I'm not dissing Paolo, I think > he's just being polite when faced with a totally bullshit situation.
Has the ALU Wiki made an enemy of a particular spammer, or has it just made a spammer compiled list of wikis?
If I was looking to be a wikid spammer, I'd type 'wiki' into Google. Conversely, I might be tempted to remove all uses of the word 'Wiki' if I was maintaining a community bulletin board.
I think that solutions which require human attention to deal with what are likely automated attacks are doomed to failure, simply because the humans will get demoralized, and the spamming programs won't.
The captcha solution suffers from the problem that it asks legitimate users to jump through hoops, and making it even fractionally more difficult to add content to a wiki is just going to reduce participation.
Is it too unrealistic or difficult for the wiki to generate a diff between the old page and the newly modified one, and run that diff through a Bayesian wiki spam classifier? Or even a program which, finding hyperlinks added to the page, loads the first one and looks for either the word 'Lisp' or for sexprs?
I find it a bit sad that, in the newsgroup dedicated to the seminal AI language, brainy people are asking other brainy people to spend their time doing such a soul-deadening classification/maintenance task.
Peter Seibel <pe...@javamonkey.com> writes: > Do you have a reference for wherever Rivest proposed replacing cons > cells with what sounds like tuples?
You might try checking "The Evolution of Lisp" by Gabriel and Steele.
"Tayssir John Gabbour" <tayss_te...@yahoo.com> writes:
> Just so you guys don't waste your time on this in vain, the ALU wiki > problem has nothing (nada, zero) to do with technical issues. Paolo > probably knows this, and suggested the only obvious solution.
More precisely, I suggested the solution that places the less--coding--burden on the site maintainer, Dan Barlow.
Cameron MacKinnon wrote: > I find it a bit sad that, in the newsgroup dedicated to the seminal > AI language, brainy people are asking other brainy people to spend > their time doing such a soul-deadening classification/maintenance > task.
Sad? I think it's just one of the many entertainments offered by venerable computing communities. Delving a little into the history of computing, and particularly Lisp, there's just a lot out there to bewilder and amuse.
Paolo Amoroso <amor...@mclink.it> writes: > From time to time, CLiki site are flooded by spam. This is currently > the case for the ALU CLiki, as you can see from the change log:
> I encourage Lispers to protect these valuable resources and frequently > "patrol" CLiki sites to remove spam: [...] > Restoring the original content is a bit tricky with the ALU CLiki, > which does not currently provide links to previous page versions. I
Here is how to do it. Suppose the spammed page is:
Paolo Amoroso <amor...@mclink.it> writes: > From time to time, CLiki site are flooded by spam. This is currently > the case for the ALU CLiki, as you can see from the change log:
> I encourage Lispers to protect these valuable resources and frequently > "patrol" CLiki sites to remove spam:
If I understand correctly, Wikis get flooded with URLs pointing to spam sites in order to make search engines increase their rating. So, just out of curiosity: what is the average amount of time for a spammed page to be effective for spammers, i.e. for search engines to actually index its content?
Paolo Amoroso wrote: > If I understand correctly, Wikis get flooded with URLs pointing to > spam sites in order to make search engines increase their rating. So, > just out of curiosity: what is the average amount of time for a > spammed page to be effective for spammers, i.e. for search engines to > actually index its content?
I suspect that there is no meaningful average - the frequency one's site is visited by search engine crawlers is likely linked to the frequency of site updates, as evidenced by HTTP Last-Modified headers. I don't have hard data for this, just a suspicion.
In article <04GdnY7YuLfI7kTcRVn...@golden.net>, Cameron MacKinnon wrote: > Paolo Amoroso wrote: >> If I understand correctly, Wikis get flooded with URLs pointing to >> spam sites in order to make search engines increase their rating. >> So, just out of curiosity: what is the average amount of time for a >> spammed page to be effective for spammers, i.e. for search engines >> to actually index its content?
> I suspect that there is no meaningful average - the frequency one's > site is visited by search engine crawlers is likely linked to the > frequency of site updates, as evidenced by HTTP Last-Modified > headers. I don't have hard data for this, just a suspicion.
Data point: I rarely (2-3x/year) update my site. Googlebot crawls it frequently (2-3x/week).
Paolo Amoroso <amor...@mclink.it> writes: > From time to time, CLiki site are flooded by spam. This is currently > the case for the ALU CLiki, as you can see from the change log:
> I encourage Lispers to protect these valuable resources and frequently > "patrol" CLiki sites to remove spam:
It looks like this may no longer be enough: spammers have figured how to add content to the recent changes page, which can not be edited later. I suggested Dan Barlow to disable editing at ALU CLiki, at least for a few weeks, and he has done that.
