On 9 Jul 2008, at 7:42 am, Andreas Tille wrote:
On Wed, 9 Jul 2008, Tim Cutts wrote:In principle, I'm more than happy to help out, but as with everything else, my time is limited. ...Sure - we talked about this at DebConf last year and I also remember the reasons you mentioned why you are not using packaged versions. It would be stupid if I would try to argue about your workflow at Sanger which has proven to be reasonable. (I'm probably biased but I think that providing software in packages and enabling users to pin on a certain version leadsdo better reproducable results than local builds.)
That only works if users have their own machines. Most of our bioinformatics goes on a couple of very large shared clusters, so pinning doesn't work. I agree with you in principle, which is why I do package stuff when I can (for example, I package some of the perl modules we need, such as AcePerl, because the version required doesn't change much)
Tim --The Wellcome Trust Sanger Institute is operated by Genome Research Limited, a charity registered in England with number 1021457 and a company registered in England with number 2742969, whose registered office is 215 Euston Road, London, NW1 2BE.