[SciPy-User] Central File Exchange for SciPy

David Warde-Farley wardefar@iro.umontreal...
Sun Oct 31 01:13:15 CDT 2010


On 2010-10-30, at 10:14 PM, Fernando Perez wrote:

> Ah, I think I misunderstood you, sorry.  Yes, using gist as the
> 'backend' could work... I guess if it's possible to use a cross-site
> authentication solution (OpenID or one of those things, I'm not very
> familiar with those tools), then one option would be to offer upon
> upload a checkbox 'Create gist for your contribution?'. If checked,
> the system could automatically create the relevant gist and display a
> prominent link to it on the snippet's page, fetch the code for display
> and download from gist (getting versioning), etc.
> 
> I don't know how hard/practical it would be, but it's certainly an
> intriguing idea.


I read this and had a slightly different idea: a mechanism for collecting gists and code snippets from sites like github. Basically a gist-indexing service. You submit a link to your gist, maybe some tags, and we fetch data about it, maybe have a fairly rapid moderation queue to make sure people aren't submitting spam.

Ideally, we'd be able to use GitHub gists, or whatever the equivalent mechanism is for several other sites (bitbucket, gitorious, pastebin.com). GitHub provides a gists API which makes machine-readable info about the gist available via JSON, so even just an URL and tags would be enough and you could fetch the rest via API.

Above all, I really like the gist.github.com model of forkable version-controlled snippets (under the hood). It would be nice if there was a software package that supported this that we could deploy on our own, but barring that, I think organizing topical gists from sites that *do* offer this model is a close second.

David


More information about the SciPy-User mailing list