I wrote what could best be described as a proto-rationalist Sailor Moon fanfic. Bear in mind that it's really old--I last worked on it around 2000 and it predates even HPMOR. It doesn't try to sell rationalism, but it has Sailor Moon do things that make sense. I never finished it but I got to the Doom Tree story. http://www.rahul.net/arromdee/fanfic.html
They would not have pages about works that were primarily sexual, because the advertisers prohibited it.
The fork decided not to bring the troper tales back, for just that reason.
It was suggested I post here, but there's a TV Tropes fork at https://allthetropes.orain.org/wiki . It uses mediawiki software and gets rid of the censorship at TV Tropes. (I suspect this one will never get rid of the strikeout tag for dubious reasons.)
Since you invoked TV Tropes, there's a TV Tropes fork at https://allthetropes.orain.org/wiki/ . It gets rid of the censorship at TV Tropes and also uses mediawiki, which makes things work better--you have real categories, it is possible to edit sections, etc.
Given that many people here are anime fans, I have an example, sort of. Back when Media Blasters had manufacturing errors that caused many of their disks to have mono audio, fans complained and Media Blasters claimed that their experts analyzed the disks and found nothing wrong with them.
Needless to say, the disks did have mono audio.
Of course, that's only sort of an example because it's just as likely that the company was simply lying about having consulted experts.
The point is not that free software programmers specifically refuse to accept wrongly indented code, the point is that they often refuse to accept code based on wholly arbitrary reasons. Arguing that indentation really isn't an arbitrary reason is fighting the hypothetical; replace it in your mind with something that is.
There's also the "we won't accept this bug report unless it fits this list of arbitrary requirements" gambit (if you actually do manage to submit the bug report following all the requirements, it will still get ignored anyway, but doing it this way they can artificially deflate the number of unfixed bugs and blame things on the user for not following the directions)
The patched version requires three questions. In fact it requires asking three people--if you only ask two people any number of questions, Random could pretend that he is the other person and the other person is Random.
First of all, you need a question which lets you determine who any person B is, when asked of Truth.
You want to ask "if B is Truth, then yes, else if B is False, then no, else don't answer". Presumably, you can't directly ask someone not to answer--they can only refuse to answer if they are Random or if yes/no would violate constraints, so this comes out to "if B is Truth, then yes, else if B is False, then no, else tell me if your answer to this question is 'no'."
Then you modify this question so that it can be asked either of Truth or False. As before, adding "tell me if the answer to Y is the truthfulness of your answer to this question" will produce a correct answer to a yes/no question from either True or False, so you get "tell me, in the case that B is not Random, if the answer to "if B is Truth, then yes, else no" is the truthfulness of the answer to your question, and in the case that B is Random, if your answer to this question is the negation of the truthfulness of your answer to this question'".
You now have a question that can determine who one person is, when asked of either True or False.
(The version with da and ja is left as an exercise to the reader.)
Use that question to ask person 1 who person 3 is. Then use it to ask person 2 who person 3 is.
If 1 and 2 give the same answer, then they are either Truth or False giving an informative answer or Random imitating an informative answer. Either way you know the identity of 3, and you also know one non-Random person. Ask this non-Random person the question to find the identity of a second person, which gives you all three.
If 1 and 2 give different answers, then one of them is Random. In that case, you know that 3 is not Random, and you can ask 3 who 1 is. You now know the identity of 1, and you know whether 1 or 2 is Random; the non-Random one told you who 3 was, so you have the identity of 3 as well. This again gives you the identity of two people and so you have all three.
your two-question solution doesn't work given the constraints imposed by Boolos
If the above is quoted correctly, the constraints imposed by Boolos are contradictory. It is not possible that 1) Random always answers truthfully or falsely and 2) Random always answers "da" or "ja". So the fact that I had to throw out the latter constraint is of no import--since the constraints are contradictory, I had no choice but to throw one out.