[Neurodebian-upstream] [Nipy-devel] Standard dataset

[Matthew Brett]

Yo,

> [cross-posting to neurodebian-upstream list]

:)

> On Mon, Sep 20, 2010 at 08:29:36AM -0700, Matthew Brett wrote:
>> I'd love to hear about standard datasets.  I've been collecting some,
>> but they're a bit scattered, with unclear licenses, most of them.
>
> Is there a public list of "free" datasets somewhere? We have started
> packaging a few:

Actually, I was just working a bit on the nipy data packages, and
wanted some help thinking about this.

It seems to me, that we need some sensible cross-platform way of
packaging data.   It would be very good if we could work out some way
of sharing packaging work with Debian too, while still being usable
for Mac, Windows, Fedora and so on.

This was the state of our thinking as of a while ago:

http://nipy.sourceforge.net/nipy/stable/devel/guidelines/data_files.html

Since then I noticed this effort by a friendly crowd:

http://okfn.org/projects/datapkg/

but I was a little intimidated by the dependencies, probably because I
wasn't paying enough attention.

I had started to think about moving this stuff into nibabel so we can
use data packages for DICOM files - they will be essential to keep the
thing robust:

http://github.com/matthew-brett/nibabel/blob/master/nibabel/data.py

I guess, what I'm asking, is, are you Michael or anyone else out there
interested in trying to work out a generic data packaging first-pass
draft thing?

See you,

Matthew