Hi Steffen,
On Sat, Nov 18, 2017 at 11:26:42PM +0100, Steffen Möller wrote:
> I just added a June 2017 PLoS One paper to DeepNano's
> debian/upstream/metadata and was a bit surprised about the comment
> "*Remark of Debian Med team:* There is no intend to keep continue the
> existing packaging since the program nanocall seems to serve the
> intended purpose better" in the bio-ngs task. I asked "git blame" and it
> blamed me - likely because of some rearrangements a couple of months ago
> because of which I had edited the line, anyway, I then removed that comment.
I somehow vaguely remember that I also got some hint about this relation
between deepnano and nanocall, but I do not remember the exact issue.
> In principle, I am all for directing users towards packages they want to
> use. After all, having a software packaged is already a selection. But
> any such direct comparison deserves a reference to a publication/URL or
> the direct notion of the reason and the version of the packages that was
> the basis of that judgement.
+1
> It would be rather interesting to think about how to formally express
> the preference of one tool over another. After all, any such preference
> typically only holds for specific lengths or reads or species or ... .
> We would have functional equivalence classes of tools so they would all
> formally fit as input to a certain set of workflows, but for additional
> reasons, part of which may be individual to the user, one tool is
> preferred over another (quality of reads/degradation, infiltration with
> pathogens, number of reads paired with maximal run time, quality, ....).
> I like it.
I think we do not (yet) have really good meens but a first step could be
that we Suggests (or even Recommends) a package and add according hints
in README.Debian.
Kind regards
Andreas.
--
http://fam-tille.de
