Le Sat, Apr 28, 2007 at 07:27:44PM +0200, Steffen Moeller a écrit : > On Saturday 28 April 2007 15:56:40 Andreas Tille wrote: > > On Sat, 28 Apr 2007, Charles Plessy wrote: > [...] > > > I think that with a size of ~ 3 Mo it is intersting to provide mirbase > > > from within Debian, but definitely this scheme is not valid for bigger > > > sizes. The human genome is definitely to be packaged as a wrapper, but > > > where to put the cutoff? > > > > Well, this might trigger the discussion about a data repository for Debian > > that pops up every six months on debian-devel mailing list. > > > > > Lastly, as all of this takes some space, I also wonder if it is > > > acceptable by policy to have the heavy data under something like > > > /usr/share/bioinformatics, in order to let sysadmins dedicate a > > > partition to this. > > One of the many projects I once started is a script to download the databases > for me. Hence, much like the MS fonts, the script should be in Debian but the > data not. We'd end up with well-described paths for the data but would not > spam Debian with too many large packages. > > With respect to the suggested subfolder I think that this should be come a > paramter of the script. /var/lib for those with recently bought computers > from the local discounter and tons of disk space working at home, those with > a network wil decide for something shared.
Hi Andreas and Steffen, For the particular case of mirbase, there is a release every quarter aproximately. I think that it makes it suitable for being directly packaged, and in therory, the data should go to /usr/share as it is shareable and static. But on the other hand, if the idea of having a common folder of sequence databases is accepted, it would be a bit stupid to keep small databases in /usr/share for the sake of it, separtating them from big genomes which would go in /var/somewhere. But the problem is that there is no /var/share, although it could be cool to have a nfs- or samba-mountable folder in local networks, from which people using modern OSes could access the data when they do not need a dedicated server to run the analysis. Alternatively, and similarly to java-package, we could have some wrappers which would produce .deb pacakges from freshly downloaded genomes. In that case /usr/share/genomics-or-whatever would be a good target again. Lastly, if we make enough progressees, maybe we can try to re-contact Bill Pearson and ask him if he reconsiders his decision about not distributing FASTA in Debian... http://lists.debian.org/debian-med/2005/05/msg00035.html Have a nice day, -- Charles Plessy http://charles.plessy.org Wako, Saitama, Japan -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
