Want to take part in these discussions? Sign in if you have an account, or apply for one below
Vanilla 1.1.10 is a product of Lussumo. More Information: Documentation, Community Support.
I want to draw people's attention to this edit, where the spammer is applying just a little bit of intelligence, including a somewhat relevant link in addition to the spam link. It's not very relevant, something that might be found by a keyword search, but it's the sort of thing that may lead to confusion if you only notice half of the edit.
Some sort of automatic warning system for the lab elves might be useful that goes off whenever somebody inserts an external link. (We could have a whitelist of sites like the lab itself, the Café, some other blogs, the arXiv, etc.) Is this feasible?
I just noticed that the munged the somewhat relevant link too; I wonder what purpose that serves.
To my knowledge, we've had 4 spam edits. The first was some time ago, the second a few days ago, then two more in the last few hours.
I like the idea of an automatic warning. Would a daily list of external links added suffice?
I notice that they did it again. Is our policy here just to remove the link (thereby creating a new edit) or to revert it?
A quick check through last Sunday's backup shows that this looks like a new problem. I may have missed the odd URL if my regexp wasn't perfect, but I didn't spot any dodgy links.
This particular spammer has struck a total of three times. Twice on localization and once at category theory (Eric spotted that one and deleted it).
Whilst this is not a major problem as yet, given the number of pages in the n-lab it's only due to the hawk-eyes of Toby and Eric that these got spotted.
Anybody have any suggestions for future prevention?
To answer my own question above, creating a new edit does have the advantage of preserving the IP address in the database. If we do decide to revert spam like this then ips should be noted first.
Having the link in the revisions is lousy advertising; search engines that respect robots.txt (which is all of the respectable ones) don't index our revisions. So removing it with a new edit should be as good a deterrent as reverting it in the database (and easier for most of us to do, especially with Rollback).
We would have to revisit this if we get so much spam that it clogs up the database of revisions (but in that case, removal vs reversion will be the least of our worries). Until then, I'd say remove it with a new edit, preferably with Rollback. Although I don't know much about how to do so, it should be possible to write a Firefox plugin to do this quickly.
I notice Eric commented on his spam-removal here. I recommend that spam-removal comments get "spam" in their title otherwise the other lab elves might miss them.
Maybe even a new category "spam"? :)
Yeah. I didn't know what to do, so I rolled it back. I'm not sure what the difference was between rolling back and just removing the spam though.
I wasn't even planning to note here when I remove spam; I did this time just to point out the above expected intelligence of the spammer.
Eric wrote:
I'm not sure what the difference was between rolling back and just removing the spam though.
What you did is what Andrew is calling ‘removing’ spam (if I understand him). The alternative is to ‘revert’ it, by which I believe Andrew means to remove the spam edit entirely from the database.
At least that's how I interpreted his comment #4, and that's how I used the words in my replies.
Jacques informs me that these particular additions are known in the instiki community and he's banned their IPs. There was another IP and checking through the database we got hit twice by that one. One was on the HowTo which Toby caught, the other was on mathematics which also Toby reverted. The second of these was a bit bizarre as it seems to have been merely the addition of a full stop - maybe they were testing our response to minor edits. Both were on Oct 22nd, by the way.
I hereby award Toby this month's Spam-Fighter prize of a plate of spam fritters.
(Prizes must be collected in person)
PS On the basis of this, I'm now in favour of simply editing out the spam rather than rolling it back, unless it is something particularly offensive that we don't want any trace of even in the database. Then it's easier to grep back through the database and find patterns that we may have overlooked.
Now ‘rolling it back’ means removing from the database? The ‘Rollback’ command on the previous version is the easiest way to simply edit out the spam.
I guess I should experiment with what rollback does before saying anything else. Given that revisions are linear, I assumed that rollback deleted all future revisions. But that might not be the case. I'll try it out on the lab elves web to see what happens.
Empty page with errant title example of compact self-adjoint opertors created by Anonymous User from IP 41.67.12.5
My guess would be that someone hit the link by mistake. I tried changing the name to example of compact self-adjoint operators to see if there was a page that linked to it, but nothing does. Still, there would be no harm in having such a page so maybe we should just leave it until someone has the time to put something sensible there!
Then I would rename it to compact self-adjoint operator as that one does not exist either, and it does not look like it is an entry which will need soon further splitting. We can have than a call to that entry from more general entry compact operator and from self-adjoint operator.
No complaints from me!
It now links to and is linked from the other two entries.
1 to 20 of 20