Skip to Content »

Michi’s blog » A vision for collaborative mathematics platforms

 A vision for collaborative mathematics platforms

  • June 19th, 2008
  • 8:53 am

Based on the extensive discussion at the Secret Blogging Seminar on tools for long-distance collaborations, Scott Morrison writes an introduction to source control with subversion for research collaborators.

In this post, Scott also offers, quite magnanimously, to setup and host subversion repositories for any mathematician who happens to want to start collaborating using subversion.

Which, to my mind, immediately prompts the question: why stop there? I’ve had ideas about setting up a free and easy to use platform for modern communication in the mathematical community before; but they were along the lines of duplicating‘s efforts; which isn’t really something that pays off on your efforts. Reading this, though, raised a new idea.

Why not setup a server – preferably with a university data center as backing – which dispenses free platforms with the following contents:

  • Source control. Preferably option on subversion, git, mercurial – or some such selection of modern and wide-spread systems.
  • Wiki, Blog, Ticket system – a wordpress installation, with LaTeX, and a Trac installation connected to the source control system in use would do quite well.
  • Heavy access control: one of the worries I hear pronounced often is that running a blog with your mathematical ideas display the ideas to the world before you get to publish them, thus risking you getting scooped on your research. This worry would be, to some extent, ameliorated by serving things optionally over https, by having a decent and robust access control system, and by having draconian privacy statements for the site administration.
  • LaTeX compile farm – why not set this thing up so that it can build your papers for you? That way we really end up building the mathematician’s sourceforge!

So, I guess this post is a call for volunteer implementers. I’ll launch the ideas in the fora I have access to – I’m headed for a faculty retreat tomorrow discussing a research project which seems to include Web2.0 for research communication as a topic, and will see if I can propose the ideas there; and it’s just the thing to discuss with the workgroup Information und Kommunikation of the DMV too. But just because I’m looking for people who want to run this in no way implies anyone reading this shouldn’t.

8 People had this to say...


Interesting topic. I’ve collected some my thoughts here (in Russian, sorry), they are primarily about wiki vs version control system for writing papers, but there are few useful (English) links at the bottom of that page.

I’m not able to be the main developer of such system, but I’d definitely like to join if someone will be gathering a team.



for some fellow students and me I developed latexki[1], a wiki for LaTeX documents (and a few feature-limited regular wiki pages) that allows online-editing the source, stores the files in a SVN repository which can also be used directly by regular contributors and has some features such as first-page-preview on the website and dependency tracking. It’s written in haskell, and you can see it live on the mitschriebwiki[2], german only though.

There is no access control of any kind via the web interface, and svn access is handed out personally, so it’s not yet what you want – but quite close, isn’t it?




A possible key to making such an effort successful would be to have a clean, easy to use, web interface for as many of the components as possible, tied together into a single package. For instance, combining your first and fourth points, there should be a webform where one or two clicks gives you a PDF of any version of the paper in the control system or a latexdiff between any two versions. (Of course robustness when LaTeXing things may be a fatal problem for this idea; a place to start would be the arXiv autotex script.)

Another possible component would be a virtual whiteboard server. Oh, and “1-click” arXiv submission.

A smaller scale starting point than the ideal of a single large university-backed server would be to create a virtual machine for Xen or VMWare that individual departments could drop into place. Something like this or thatbut with all the features you’re suggesting.


You should look into some of these sites set up for programming projects. At the moment, I’m trying out assembla, which has a lot of the features you want (Subversion, Trac, wiki), though obviously not so much the LaTeX support.

  • Michi
  • June 19th, 2008
  • 19:47

Ben: As you might have noticed above, I draw my inspiration solidly from the programming project sites – such as SourceForge, Rubyforge, assembla, ….

Thing is, these don’t have LaTeX support. The non-plus-ultra of mathematical collaboration. They also aim programmers, not mathematicians. Which is why I think that we need a site for our needs.

  • Nathaniel Dean
  • June 19th, 2008
  • 21:02

It’d be really cool if you could integrate Coq or other proof verification tools into the setup. That way you could even run “unit tests” on your paper for a sanity check.


This sounds like a great idea. I’ve often felt that mathematicians are kinda behind the curve with respect to good use of computers and the internet — the arXiv was started by physicists; darcs, with its “theory of patches” and “patch commutators”, was started by a physicist…guys, we’re getting pwned by physicists! :)

Speaking of physicists, this site would be very useful to them, as well as computer science people, and anyone who needs good collaboration with LaTeX documents. The potential audience, I think, is pretty big.


Hi, I’ve been wishing for some sort of online LaTeX collaboration for a long time as well, so I ended creating a website to try provide some of the features you mention. It started out mostly for myself, but now I’m hoping other people might find it useful. The site is and it renders LaTeX, has a revisions system and provides comprehensive access control. I’m hoping to add bibtex support in the near future.

Want your say?

* Required fields. Your e-mail address will not be published on this site

You can use the following XHTML tags:
<a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>