I am not very familiar with Tricki, so I hope I am not repeating a discussion that has already occurred. But I have an idea that I think would do the Mathematics community a lot of good. I am not sure that Tricki is the best place to implement the idea, but I'd like to see what you all have to say.

As a graduate student, I know I have poured over many papers for extended periods of time, trying to understand a bit that is not written very clearly, or has some missing details, or that just went over my head. Papers are written for different audiences, and sometimes I am not a member of that audience. Often times I will work through some of these details to convince myself that something is true. I am sure we have all done this many times over.

Suppose we created a wiki organized by PAPERS with comments specific to the paper, maybe even paragraph by paragraph. If you have spent a lot of time with a particular paper (not your own), and think you have a way of explaining a point in a way different from the author, it could save another several hours to include your observation on a collective comments page. This could also be a way of making errors (severe or trivial) more widely recognized, instead of one needing to know that there is an errata page located somewhere on someone's website.

This is just a thought I had, and I was curious about the kind of support or interest there may be in such a project. It seems to me like it could be a natural extension to something like Tricki, or maybe even the ArXiv. I welcome comments and further discussion about this topic.

I think this is a great idea and I wonder if the arXiv would do it. It would be as simple as attaching a comments section to each paper's page, below the abstract. If arXiv won't do this for some reason, it would suffice to have something that searches the arXiv and keeps comment pages for each paper.

Incidentally, did you get this idea from a recent PhD comic? :)

That does sound like a good idea. My worry would be that it is clearly open to abuse, and so that would need to be controlled in some way.

I like your idea. This is also good for centralizing folklore errata for papers and textbooks.

Somebody could write "I think Proposition 2.1 is missing a compactness assumption on space X. But I can't find an example to demonstrate that. I'm not familiar with this field." and then somebody else could confirm it.

Somebody could write "Corollary 3.1 is just false because ..." and then somebody else reply "It's actually well known in folklore that it is false."

This is something I have been thinking about for years.

- The comments could also be pointers to newer articles that extend a specific theorem, clarify a certain passage, disprove a statement, etc

- The comments could also be "details left to the reader", submitted by the author her/himself. Oftentimes I leave out some computations before submitting my papers, for length considerations.

It would also give a more interlinked structure to the arxiv, allowing maybe one day for more intelligent harvesting of the mathematical literature.

How about using Google's Sidewiki for this purpose?