[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Data packages.

Le Mon, Feb 09, 2009 at 12:57:12AM +0100, Peter Palfrader a écrit :
> The data.debian.org thing is actually Joerg's proposal, he first raised
> it at http://lists.debian.org/debian-devel/2008/05/msg00970.html last
> year.

Hi Peter,

thanks for the pointer.

I think that I will keep on exploring the possibility of distributing data in
Debian binary format with no Debian source package for the moment.  Jorg's
proposal just doubles the bandwith requirements for the uploads, and many
databases are updated monthly… I plan to use our getData tool to update some
mirrors locally, and build Debian packages from this. By the way, I am
wondering about the possiblity to use hard links during this process to avoid
unnecessary duplication of large files. Any hint ?

My current plan is :

 - Prepare a Debian Med image for the Amazon Elastic Computing Cloud, that
   would only contain programs packaged in Debian, with the exception of
   packages from backports.org, which I think are trustable enough even if they
   are not official.

 - From within the Amazon system, build binary packages of bioinformatical
   data, and distribute them on the Amazon Simple Storage system, if it is
   possible to open to the inside (free transfer), but not to the outside (costs
   me money).

 - See if people use the data packages in conjuction with the
   unofficial-but-gpg-signed Debian Med image that I indend to prepare.

 - Ask for sponsorship if it starts to cost too much :)

Of course, help from people interested is most welcome! There are a few started
threads on the debian-med mailing list, for which I have not found time to
answer properly because my lack of progress: in parallel to this project there
is a major update of BioPerl, that pulls many things by spaghetti effect. But
be sure that I welcome all feedback!

Have a nice day, 

Charles Plessy
Debian Med packaging team,
Tsurumi, Kanagawa, Japan

Reply to: