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.
This is new: when trying to save an addition to the entry Chern-Simons theory I get the message
Your edit was blocked by spam filtering
In case anyone else can more successfully pretend not to be a spammer and wants to help, this here is what I was about to add to the References-section, at the end of “With Wilson loops and defects”:
Discussion of boundaries (branes) and suface defects for Chern-Simons theory is in
Anton Kapustin, Natalia Saulina, Surface operators in 3d TFT and 2d Rational CFT in Hisham Sati, Urs Schreiber (eds.) Mathematical Foundations of Quantum Field and Perturbative String Theory (schreiber) AMS, 2011 {#KapustinSaulina11}
Anton Kapustin, Natalia Saulina, Bicategories for boundary conditions and for surface defects in 3-d TFT, Nucl.Phys.B845:393-435,2011 (arXiv:1203.4568) {#KapustinSaulina12}
Jürgen Fuchs, Christoph Schweigert, Alessandro Valentino, Bicategories for boundary conditions and for surface defects in 3-d TFT, Hamburger Beitr. zur Mathematik Nr. 433, ZMP-HH/12-5 {#FSV}
Hm, I can still edit other entries. So maybe some substring of the above text triggered the spam filter?
Ah, I have isolated the problem a bit. It is this string of symbols here which causes the spam filter trigger:
Hamburger Beitr. zur Mathematik Nr. 433, ZMP-HH/12-5
okay, I get it. The offensive term is
Hamburger
Hm, now that’s some example of a global misunderstanding.
Reminds me of JFK. Ich bin ein Berliner = I am a doughnut. Then Ich bin ein Hamburger, ich bin ein Frankfurter, ich bin ein Wiener … you get the idea.
And not to forget: Ich bin ein Amerikaner.
(I should check if that triggers the spam filter… ;-)
Good one; I didn’t know about that! Here in the States we just call that “the black and white cookie” (as made famous in an episode of Seinfeld).
So the filter is confusing spam with hamburger? They're both cheap meats, I guess.
…. but what type of meat!!!! Spam is I believe distantly related to Ham (It was ‘spiced ham, according to its inventors) Hamburgers are very distantly related to ….??? Horsemeat??? and on looking up the word in Wikipedia it is the bun not the meat that is really the Hamburger. Confusing, but consistent with doughnuts and other earlier comments.
In the U.S., ‘hamburger’ may be used (albeit improperly) as a synonym for ‘ground round’, which is the meat that we make hamburgers out of. And hamburgers without buns are common too; we rarely call them ‘patties’, despite what Wikipedia implies. There is even a brand name, ‘Hamburger Helper’, for boxed meals of noodles and dried sauce/gravy, to which one is expected to add ground meat, not a sandwich.
While this discussion is wandering off into the direction that it is wandering off to, for the record I’ll just recall that implicit or not so implicit in the above is the suggestion that it would be nice if on the Lab one could mention the mathematical preprint archive of the University of Hamburg called, unfortunately but not unreasonably, Hamburger Beiträge zur Mathematik.
Why not abbreviate in some way so that the first word doesn’t trigger unwanted effects? Hamb. Beit. zur Math., with a link perhaps?
I’ll take a look at the patterns to see what’s possible.
I don't understand why that word would be flagged anyway. The meat connection notwithstanding, hamburger is rarely marketed through spam.
hamburger is in the list of spam patterns. I could easily remove it if so desired.
(I don’t know where the list of spam patterns originally came from; my guess would be that it was in instiki before Jacques started working on it.)
I would desire that. I understand why it used to block ‘specialise’, but this entry makes no sense. We can remove it safely, and you never know when more research will come out of Hamburg.
I’m getting this same message “Your edit was blocked by spam filtering” when editing at 11-dimensional supergravity
The basic BPS spates of 11d SuGra are
I wonder what’s so troubling about BPS states.
That is very strange. Not only did I get the same as one would expect, but when I copied the line then deleted and pasted it back in again (with ‘spates’) it gave me the same result although there was no change to the text. (Is there a special character hidden somewhere?). The page BPS states loads without problem and doing a minor edit did not produce any reaction. (I then deleted the minor edit!)
On it. Same problem as here.
Fixed now, by adding a small exception to the regex in the spam filter that I added to deal with the persistent problems that a certain user has been causing; it should still be tight enough, but will hopefully now not exclude most typical cases. It is conceivable that certain arXiv references or other references, and things like IP addresses, could still be problematic; let me know if you come across an occurrence of this, and I’ll try to address it. As mentioned at the other thread, I would like to keep the filter as tight as possible for the moment.
For future reference, the way the spam filter works is that it scans the entire page to make sure there are no occurrences of certain words/patterns. In this case, the bit that you were trying to edit, David, had nothing to do with the reason the spam filter blocked the edit; there was an offending pattern somewhere else in the page.
I see. Thanks.
Thanks for catching the spates! And thanks to Richard for fixing the issue with the spam filter!
Hi jmoeller, I’d suggest to add the text in smaller pieces, or else paste the full source here and ask one of the ’trusted users’ (those who have made many edits, etc) to do it. Apologies for the inconvenience, but we need this kind of blocking in place!
By the way, one of the spam detection metrics is to match the nLab author with an nForum user. The algorithm for this cannot match ’Joe M’ with your nForum user; I’d suggest to sign your nLab edits with your full name.
I just wished to observe, since there are occasional instances where the spam filter still blocks valid edits as in above, that I just checked the logs now, and it has blocked several instances of spam that would previously have gone through. Indeed, it is blocking more things that it should be blocking than it should not be blocking. So I think it is roughly speaking doing its job.
I cannot bring this addition (mainly references to related entries and an external resource) to the so far essentialy empty catamorphism page past the spam detector, even in parts:
Given an endofunctor , the catamorphism for an -algebra is the homomorphism from the initial -algebra (if existent) to .
I’ve inserted the text at 29 into catamorphism.
From #30:
The trick is to start with very small pieces of text at the beginning, such as only a phrase or two (so not even a full sentence) so that the edits get past the spam filter. Then, one would open the editor again slowly add more text, each time increasing the amount of text placed in the article. Eventually it snowballs and the article should function as normal.
Thanks for highlighting, I had not been aware that the behaviour of the spam filter is that unreasonable. I’ll try to bring this up with the technical team to see if they can do something to improve the situation.
Happy to announce that the spam filter has been adjusted (thanks to Christian Sattler!):
It’s less strict on the edit threshold now and throws a slightly more informative message when triggered.
By the way, the edit threshold is (and currently remains) in proportion to the ambient page size – for better or worse. This means that the unreasonably strict trigger threshold as reported in #30 will still apply to the editing of pages which are almost empty. Hopefully this is a rare situation.
By the way, the edit threshold is (and currently remains) in proportion to the ambient page size – for better or worse. This means that the unreasonably strict trigger threshold as reported in #30 will still apply to the editing of pages which are almost empty. Hopefully this is a rare situation.
From experience, it really isn’t.
There are a lot of very short stub articles on the wiki, so trying to expand the stub articles into full-fledged articles would result in triggering the spam filter if one isn’t careful about how much text one adds to the article.
When trying to merge and redirect entire articles into other articles, one has to remove the text in sections, or otherwise the spam filter would trigger.
Well after the spam filter changes, mergers became a lot easier than before.
re #34: Yeah, I can see that it’s a pain. Sorry about that.
We talked about removing the spam filter altogether, but it seems dangerous at the moment. Ideally it would probably be good to instead have a completely different scheme where edits on the nLab happen through verified password-protected accounts. But it seems we need to wait until we have hired a dedicated sysadmin to implement such more far-reaching changes.
However, the new spam filter configuration should now also have a lower value of page edits for a given user from which on the filter stops interfering. So if,as it sounds, you are already an experienced editor, then you are probably close to not being bothered by the filter anymore.
1 to 36 of 36