Start a new discussion

Not signed in

Want to take part in these discussions? Sign in if you have an account, or apply for one below

Site Tag Cloud

Vanilla 1.1.10 is a product of Lussumo. More Information: Documentation, Community Support.

1. removing category: World [private]

Valeria de Paiva

• CommentRowNumber2.
• CommentAuthorUrs
• CommentTimeNov 4th 2020

What a weird page. I have removed some more weirdness.

I don’t think it was me who created this, contrary to what the page history shows.

• CommentRowNumber3.
• CommentAuthorRichard Williamson
• CommentTimeNov 4th 2020
• (edited Nov 4th 2020)

It dates from the very earliest years of the nLab. It looks like it might possibly have been a page on your personal web or something once; maybe it was moved accidentally to the nLab during some early software changes or something. There are one or two other weird pages from the early years lying around, which I sometimes discover when working on the software, because they often lead to ’corner cases’; I usually tidy them up when I see them, if I can.

Anyhow, the good news is that with the upcoming nForum and nLab code rewrites, logging in will become possible on the nLab, which will protect usernames from people masquerading as others, just in case that was what happened here (although as I say I think it is doubtful in this case).

• CommentRowNumber4.
• CommentAuthorUrs
• CommentTimeNov 4th 2020

I know, we had seen more blatant such entries before, allegedly signed by me. I am trying to entertain the idea that my memory was completely blanked of systematic attempts that the record claims I made 10 years ago on adding, besides some nonsensical phrases and category-labels, a generic list of search engine links to entries, some of which I never use and some of which I hadn’t even heard of back then.

To me, this strange stuff looks like an attempt of a script to add generic pieces (section headers, categories, generic search engine pointers) to entries.

I find this decidedly scary.

• CommentRowNumber5.
• CommentAuthorUrs
• CommentTimeNov 4th 2020

Just look at the other example that Valeria found:

On Aug 26 2011 I announced having added “some references” to Morse homology.

The source of that revision shows three references like me, a human, would add them, interspersed with a good bit of whitespace, as is my habit.

Then right below that, not separated by any whitespace, is the exact same generic nonsense content that we also see in the history of Riemannian cobordism, just like an automated script would add it.

Had I added that content, why would I not have expanded the announcement by “oh, and on top of that I added a list of generic pointers to search engines, and a nonsensical category-label and, last not least, the comment that there is also an nLab page with the same title as this nLab page that we are on, just to prove that I’d fail a reCAPTCHA test”.

Therefore, even assuming that I am suffering from abysmal amnesia combined with split personality syndrome, I don’t see how I would have typed that material, or how any non-scripted entity would have. Of course the conclusion of that is no less scary.

2. just like an automated script would add it

Yes, that this was some weird automatically generated content seems the most likely explanation. Probably Andrew Stacey would be the best person to know if there was some kind of experiment with that at that time on the nLab server.

Anyhow, what I do know is that there are fortunately not many pages like this, only a handful of outliers; I don’t think there has ever been any wide-scale masquerading as others. Thus I wouldn’t worry about it too much, I think everyone is aware that you would not write this kind of nonsense.

But definitely the best solution in the long-term is to be able to log in. That functionality is basically written for the new nForum and will be made live as soon as I get a chance; after that it should be easy to adapt it to the nLab.

3. If you prefer, I could edit those dodgy revisions in the database to cut out the bits which are obviously ’non-human’, so that your name is not associated with them?

• CommentRowNumber8.
• CommentAuthorUrs
• CommentTimeNov 4th 2020

Hi Richard,

thanks. I wouldn’t mind, but it’s not important.

After all, despite looking weird, these ghost edits do not seem to be intended to do harm; they appear to be an ill-conceived attempt at being useful.

Which makes me think that the intent was not to masquerade under my name. Instead, the most plausible scenario I see is that somebody with root access was writing into existing entries without triggering a new revision, and that these previously existing entries all ended up being entries signed by me just by chance, by the law of large numbers.

This would also explain why nobody noticed anything strange when I announced the edit Aug 26 2011 (at a time when the nLab was much smaller, when all edits would attract attention from regulars): Maybe there was nothing strange on that day, and maybe the weird code was added only later by root activity, unannounced and without generating a new revision number.

• CommentRowNumber9.
• CommentAuthorRichard Williamson
• CommentTimeNov 6th 2020
• (edited Nov 6th 2020)

Which makes me think that the intent was not to masquerade under my name. Instead, the most plausible scenario I see is that somebody with root access was writing into existing entries without triggering a new revision

Yes, I agree, something like that is probably the case.

I have now tidied up the history as promised, both on this page and at Morse homology, by removing the ’generated’ content, which means that Valeria’s #1 now shows up as not making any change.

• CommentRowNumber10.
• CommentAuthorRichard Williamson
• CommentTimeNov 6th 2020
• (edited Nov 6th 2020)

Same for K-cohomology and L-theory. See also the discussion here from 2018 on this.

Hopefully all cases of this have now been handled.

• CommentRowNumber11.
• CommentAuthorUrs
• CommentTimeNov 6th 2020

Thanks, Richard!