I think this question should stay open. It's something mathematicians would be interested in knowing, and I'm not worried about questions like this one taking over the site.

I was undecided about voting to close, but ultimately I think that I would not want to see this sort of question on MO. Even if it is of interest to some mathematicians, it is ultimately not about mathematics. I too have voted to close.

Noah: I am. Not about djvu in particular, or even about something a little more relevant to mathematicians' daily lives (like LaTeX questions), but questions where there is no expert present. Or rather, where there can be no expectation that an expert is present. There may well be many people on MO who know about djvu or LaTeX or bus routes through London, but if I ask about the best way from Hammersmith to Mornington Crescent then even if someone gives me an answer, I cannot have any expectation that their answer is going to be reliable.

Mathematics questions are completely different. As a mathematician, I know how to assess whether or not to believe another mathematician. I have lots of tools available to me and lots of ways to check any information that I receive here. So I can freely ask, knowing that I can assess the reliability of the answer. Indeed, unless I know before I ask how I will verify an answer then I should not ask it. This question does not fit that, and so - in my most humble opinion - it does not belong on MO.

That's my ultimate dividing line. As MO is for experts, the questions should be those that are answerable by those experts in the subject in which they have expertise, not in random snippets of knowledge that they might have picked up here or there.

(Incidentally, I see that I have failed to obey the protocol: I did not leave a comment on the original question linking to this discussion. I apologise for that omission.)

I wish this question had not been closed. In my opinion this question is of interest to a number of mathematicians and is also a question that mathematicians may have specific knowledge about.

Many mathematicians, especially younger ones, have an interest in making all mathematics books available in digital form; for evidence just witness the vast number of mathematics books that have been scanned and are available at the present moment. These large collections of scanned mathematics books were scanned by mathematicians themselves---no one else had (or has) any incentive to sacrifice the considerable amount of time needed to scan these books. Moreover, if you take note of the university insignias that are stamped on some of these books you might notice how wide-spread this work is----these scanned books are coming from the libraries of universities all over the world. Currently DJVU is the best format for these kinds of scans of old books and many (perhaps most) of these scanned mathematical libraries are in DJVU format (rather than say PDF).

For these reasons I feel that questions about dealing with DJVU documents, including the knowledge of how to create them, is more relevant to mathematics than to a generic point of computer scientists, or other computer types. Or to put it differently, mathematicians have more of a reason to have such expert knowledge (as they are the ones scanning these mathematical libraries) than random computer programmers (especially given the non-standard nature of the DJVU format).

Should we be discussing how to deprive authors and publishers of possible sales?

Maybe publishers, because I don't think that authors of the maths books (budding) research mathematicians would be interested in stand to make a lot of money from their books.

But I think that this is not the point. I think that the question is not ultimately mathematical, and this is why in my opinion it is not appropriate to MO.

Should we be discussing how to deprive authors and publishers of possible sales?

The way that books at the graduate level and beyond are published is pretty rotten. They offer a laughably small amount of the profits to the authors as royalty payments. It seems like a holdover from the days before the internet. The only books I would ever consider buying are those from the mathematical societies (AMS, EMS, SMF, etc.) or university presses, since at least the money is going to a good cause. Well, I guess I've bought books from springer a few times, but only because it's Bourbaki.

Before this gets off-topic, it was my intention to make a specific point that "knowledge about how to create DJVU files" is knowledge that we might expect many mathematicians to have expert knowledge about (because of their careers as mathematicians). I do not agree with jbl's take on my argument. I also do not see how the ethical issues that might be related to my examples are at all relevant. But in any case, for an example which I doubt any people would have moral hiccups about, take this one: In the past I have asked a colleague at a different university to see if they can get a paper out of a journal that my university's library does not have and if they could scan it and email it to me. They did this. They turned it into a DJVU file and emailed it to me. I would bet that the only DJVU files my colleague has ever made were from scanning a paper or section of a math book in this kind of context---that his knowledge on how to create DJVU files has come about specifically through his work as a mathematician.

Anyway, I don't think it is a big deal. I just noticed that many people think that this question is not good for the basic reason that jbl and others have put forward. I agree that that is a reasonable viewpoint in general, just one I don't agree with in this particular instance. To me, this is a question that is definitely of interest to mathematicians as mathematicians, the answer of which requires knowledge that mathematicians may have because they are mathematicians. [That's a lot of mathematicians in one sentence!]

Publishing as a trade is just barely above the protection racket level.

Well said. Is there anything that could be done?