From: Geoffrey G Thomas Date: Fri, 2 Oct 2009 19:22:20 +0000 (-0400) Subject: add everygit idea. X-Git-Url: https://sipb.mit.edu/gitweb.cgi/wiki.git/commitdiff_plain/3a43fabb08e6269cd234c0139343ad4a720ece83?hp=e24b4854f06186774bd73757c97d05c1b86683c9 add everygit idea. --- diff --git a/projects/ideas.mdwn b/projects/ideas.mdwn index 1375376..dcfa9f4 100644 --- a/projects/ideas.mdwn +++ b/projects/ideas.mdwn @@ -168,6 +168,25 @@ database. One possible platform is the [Exhibit](http://simile-widgets.org/exhib _Contact: pbaranay, fawkes_ +## Search across every public git repository + +The [git](http://www.git-scm.com/) version control system has been +rapidly gaining in popularity both in the world of free software at +large as well as at here at SIPB. One of the properties of git is +that every object, whether it's a specific version of a file or a +changeset to a project's source code, has a unique identifying hash +(e.g., "8c5fa44b1a4902ba1f84a71df20bfe4f21867a7c"). Given one of these +hashes, it should be possible to identify which project it came from +and what it refers to. + +The "easy" way to do this is to create a massive single git repository +with references to every public git repository we can find, but that +probably causes performance problems, so you may need to be cleverer +about how data is stored and expose a different interface. You'll +also need to figure out a way to find all public git repositories. + +_Contact: geofft, nelhage_ + ## Your Project Here SIPB can help you out in terms of both computing resources and