On Sun, 14 Feb 2016, Ole Streicher wrote: > >> There are some problems why Debian is not soo widely used in the science > >> analysis (yet): One is that due to our long freeze many important > >> packages are already outdated when the stable release comes out.
> > Well, they're using oldoldstable and old^3stable, so it doesn't appears as > > if they rely on newest versions of packaged tools. > It probably varies with the group. I discussed about the inclusion of > the analysis suite, and there the argument was that it would be outdated > in stable already when it comes out. FWIW, as some might not know, in NeuroDebian project (http://neuro.debian.net) we are consciously trying to package any of software we maintain so backport builds can be done automagically without major additional investment of time with a helper script, see for that https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=660208 and up-to-date version is always available from e.g. http://anonscm.debian.org/cgit/pkg-exppsy/neurodebian.git/tree/tools/backport-dsc This way we provide most recent releases of software we maintain at once not only for Debian sid, but for all compatible older Debian and Ubuntu releases, see e.g. for our own project: http://neuro.debian.net/pkgs/python-mvpa2.html I guess any other blend/team could adopt similar strategy. And whenever Debian "PPA" efforts comes out, it could/would be fully automated. > Data analysis is here probably different from storage clusters or web > servers. ;) I would say that "it is easier", since usually those (analysis suites) are leaf packages and thus there is fewer of dependents, so providing a new backported version is of lower risk to brake anything. What I would encourage though is to enable build-time unit-testing. It was of paramount importance/help to us at least, since then we catch possible incompatibilities at build time and never ship a backport which we know wouldn't work on older system. </shameless-plug> Cheers, -- Yaroslav O. Halchenko Center for Open Neuroscience http://centerforopenneuroscience.org Dartmouth College, 419 Moore Hall, Hinman Box 6207, Hanover, NH 03755 Phone: +1 (603) 646-9834 Fax: +1 (603) 646-1419 WWW: http://www.linkedin.com/in/yarik
