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

Re: Before I send a bug report to change DebTags please do some sanity check



Hi Enrico,

On Mon, Feb 08, 2016 at 11:09:16AM +0100, Enrico Zini wrote:
> 
> I had a look at the diff[1].
> 
> Biology is not at all my field, and the "biology" facet is clear and
> bounded enough that I have no objections to pretty much any change in
> it. Feel free to take care of it as you wish and commit changes to it
> directly to master.

OK.  So I will discuss these changes with the biology experts in detail
(since I also do not consider myself a real expert.)
 
> I'm also entirely in favour of the restructuring of facet to remove the
> tags under biology, that I agree belong in the biology facet.

Sorry, I do not understand this sentence.
 
> I have a problem with having a new "edam" facet,

OK.  Thanks for your opinion which is really welcome since I did not yet
developed a proper feeling what might be a sensible facet and what not.

> and with very
> field-specific formats in "works-with", because I see them as very
> specific concepts creeping into a much more general scope, as if
> everyone using Debian was supposed to have an interest in biology.

I guess this was the initial motivation to rather have

    biology::format::*    instead of    works-with-format::*

Is this correct.  I'm fine with reverting the change.  I was guided by
the consideration that everything that has (could have) a mime type
might be specified as works-with-format::*.  The fact that there is an
ongoing effort to register mime types was motivating me for the change.
Is my general consideration that there should be a mime tyoe
works-with-format:: match sensible or not (and should we possible
refresh the discussion in case there might be mime types for biological
formats or not).

> My problem is probably triggered by a general discomfort that I
> accumulated over the years by seeing the flat Debian package namespace
> being slowly and steadily polluted with very specific packages with very
> short or very general sounding names. None of these packages does
> anything remotely similar to what their names suggest to me: wise,
> velvet, treeviewx, t-coffee, raster3d, readseq, seaview, plink, muscle,
> melting, loki, kalign, infernal, gmap, garlic, fastlink, exonerate, bwa,
> bedtools, autodock, arb, alien-hunter.

I perfectly agree and I also do not like this.  While I could somehow
live with the naming of the packages which is rather a matter of
aestetics in my eyes the polution of the name space in the file system
is really annoying and leads sometimes to real problems.
 
> I do not know how to fix this at a package namespace level, and I would
> like to avoid the same to happen at a tag vocabulary level.

Agreed.

> I understand that, for people working all day with bioinformatics, all
> those things are as common and as everyday as water and air, and I don't
> mind if in the system that bioinformatics people use they are treated as
> such.
> 
> In a general, universal context, however, I'd like to avoid toplevel
> creep of field-specific concepts, and try to design things so that
> concepts from all of the fields where Debian can possibly be used[3] can
> possibly coexist, and not compete on who's the first to allocate a three
> letter acronym on a flat namespace.

If you ask me for my personal opinion I would not insist on this first
comes policy if there are good reasons to move a biology specific
program from the generic name space.  We did this sufficiently enough
when coming as second and have worked out some workarounds.

> How about biology-edam facet, and a biology-format facet?

Just to be sure:  With this syntax you suggest really new facets and not
tags inside the biology facet, right?
 
> We could also decide that each tag under "field" can act as a prefix for
> any number of field-specific facets, maintained by the people who
> package field-specific packages.

Sounds like a good compromise to me.

> What do you think?

I wrote to you in the first place since I expected a sensible and
helpful answer and I thing this fits exactly my expectation.  I'd wait
for further clarification where I was not fully sure whether I
understood correctly.
 
> [1] http://anonscm.debian.org/cgit/debtags/vocabulary.git/diff/?id=edam&id2=master
> [3] say, for example, Arts, Astronomy, Aviation, Biology, Chemistry,
> Computer Science, Electronics, Financial, Genealogy, Geography, Geology,
> Linguistics, Mathematics, Medicine, Meteorology, Physics, Religion,
> Statistics, and more.

Any other opinion?

Kind regards

     Andreas.



-- 
http://fam-tille.de


Reply to: