Bugzilla Bug 138049

Looks like it is a time to create a bug to track reorg progress. I will post
the proposed list (so far only 1 reject to the list of involved people) here.
If there is an expected non-trivial discussion of some particular
category/herd, please create a separate bug and make it blocking this one, so
that we can stay organized..

Proposed herd names have sci- stripped, but I am not sure how to go about that
really atm..

The preliminary split:
category: #packages (proposed herd name), comment if any

sci-astronomy: 11 (astronomy)
morfic, phosphan, zx, ribosome, aliz, corsair, mr_bones_

sci-biology:  58 (bio?)
ribosome, corsair, j4rg0n, mcummings, sediener, pbienst, apokorny, 
hansmi?, phosphan, lostlogic?

sci-calculators:  26 (calc?)
centic?, cryos, ribosome, spyderous?

sci-chemistry:  50 (chem?), sci-crystallography was proposed to split off
spyderous, markusle?, phosphan, marcus, hannes, ribosome?, kugelfang, 
agriffis, hansmi?

sci-electronics:  34 (electronics?)
calchan, chrb?, agriffis?, phosphan, ribosome, plasmaroo, hansmi, 
cryos?, gustavoz?

sci-geosciences:  8  (geosci?)
ribosome, spyderous, cryos, nerdboy, mholzer?

sci-mathematics:  34  (math?), interest in sci-proof was mentioned, possible
split in symbolic and numeric..
plasmaroo, agriffis?, mattam, cryos, ribosome?, markusle, spock, phosphan

sci-misc:  19 (sci)
cryos, hansmi?, phosphan, ribosome, kugelfang?, pbienst

sci-visualization:  20 (??)
markusle, phosphan, ribosome, cryos, kugelfang, latexer?, j4rg0n?, corsair?, 
spyderous


Looks like we can have new categories: 
sci-crystallography - seems to be enough packages and maintainers
sci-mathematics may be split into symbolic and numeric (anybody to really
comment?).

Few more categories-to-be if there is enough "material". May start as herds
under -misc or something more related:
sci-cad (cad?)
sci-physics (physics?)
sci-linguistics (??)

George

A short note:

It may not be worth trying to get it all done at once - this is mostly
"internal stuff" and I am not going to push anybody who is not going to put 5
minutes into thinking about it. However lets at least split off the ones where
there is an interest. So, the procedure could be:

1. Create a bug for the herd/new category, listing packages and involved devs.
2. Let people comment - would one week be enough
3. Create a herd or category (if there are enough packages) and adjust
metadata/move stuff around..

How about this approach?

The "to be" herds that have seemingly shown enough interest:
crystallography,
sci-math-proof,
electronics
?physics?

George

Ok i'll take a look at some physics related packages and start a new category
for it if it's needed. I'm offering help to take care of physics packages and
start a "sub-project" for physics if noone is against it.

Herds I should probably be in: chemistry, biology, crystallography,
visualization, maybe physics. Not calculators or geosciences.

I would definitely like to contribute to sci-physics.
Otherwise I'd like to be in chemistry, biology, and visualization since this
is where my personal and professional interests are.

Been really busy as of late, so haven't contributed as much as I might have to
the discussion. The marcus dev is actually me when the environment variable
wasn't set for a few months... Doh! So marcus = cryos. I can see the point to
splitting off the sci herd if people really are getting worried about its size
and it is scaring them away.

I have been very busy lately and so Gentoo has not received as much time as
usual. I would definitely support a sci-physics herd and category if there were
the packages. I would also like to be in it. I would join astronomy (maybe),
calculators, chemistry, electronics, geosciences, mathematics, misc and
visualisation. I would be in favour of shorter names, but can live with the
longer ones. I think sci-crystallography would arguably work well in
sci-physics, but if not I would probably join sci-crystallography too as I do
work with X-ray/neutron diffraction and reflection.

I think we should be a top level project though and would fully support that.
Our herd testers and overlay are doing pretty well too. Hoping to catch them on
IRC soon too.

(In reply to comment #5)
> longer ones. I think sci-crystallography would arguably work well in
> sci-physics, but if not I would probably join sci-crystallography too as I do
> work with X-ray/neutron diffraction and reflection.

As a protein crystallographer, I simply can't support this. =)

If there is something related to physics imo this are the packages i saw:
please add any i missed

/sci-misc/camfr
sci-misc/lightspeed
sci-misc/mpb/
sci-misc/oww ?
sci-misc/xfoil
sci-astronomy/maestro
sci-chemistry/shelx
sci-chemistry/rasmol
sci-chemistry/probe/
sci-chemistry/platon
sci-chemistry/ortep3
sci-chemistry/molmol
sci-chemistry/moldy/
sci-chemistry/molden
sci-chemistry/mead/
sci-chemistry/maid
sci-chemistry/eden
sci-chemistry/cns/
sci-chemistry/abinit

a few sci-chemistry packages like molmol is used in both chemistry and physics
(for solid state).

Herds I should be a member of: biology, chemistry, calculators and mathematics.
Not astronomy, electronics or geosciences.

(In reply to comment #6)
> (In reply to comment #5)
> > longer ones. I think sci-crystallography would arguably work well in
> > sci-physics, but if not I would probably join sci-crystallography too as I do
> > work with X-ray/neutron diffraction and reflection.
> 
> As a protein crystallographer, I simply can't support this. =)
> 

I think crystallography should be placed as a new top-level sci-category
(sci-crystallography), as it does not belong to either physics, chemistry,
biology or math strictly, but all of those use and contribute to this special
science.

(In reply to comment #7)
> If there is something related to physics imo this are the packages i saw:
> please add any i missed

> sci-chemistry/shelx
> sci-chemistry/rasmol
> sci-chemistry/probe
> sci-chemistry/platon
> sci-chemistry/ortep3
> sci-chemistry/molmol
> sci-chemistry/moldy
> sci-chemistry/molden
> sci-chemistry/mead
> sci-chemistry/maid
> sci-chemistry/eden
> sci-chemistry/cns

I will not support moving any of the above to physics, although I would support
sci-crystallography for those related to it.

> sci-chemistry/abinit

This should probably be in a sci-physics, as well as the CAMPOS ebuilds in
Bugzilla somewhere.

i understand but i use molmol and molden for studing element structures in
solid state physics. This is a little difficult to organize since yes it could
be on crystallography

(In reply to comment #11)
> i understand but i use molmol and molden for studing element structures in
> solid state physics. This is a little difficult to organize since yes it could
> be on crystallography

We could consider moving them to sci-visualization, but they seem good enough
where they are and it's more valuable to me to retain the CVS history.

I could contribute to sci-physics, sci-astronomy, sci-libs and some dev-python
packages related to science. My specialty is cosmology and high energy physics.
I did some ebuilds in the scientific overlay (as bicatali), and has contributed
some bugs, as a sci herd tester. By the way, doesn't the scientific overlay,
which was also created for beginners to train on package maintaining, overlap
with the mentoring process?

Sebastian please add yourself to bug 138711 so we can discuss a bit.

As Sebastien, I'm a herd tester. I would like to contribute to sci-electronics
and sci-mathematics mainly, and to sci-visualization and sci-physics in a
smaller amount.

Thanks,

Lucas Chiesa (tulku)

Everyone that want's to be on sci-mathematics herd please join bug 139682. For
Herd Testers after get this reorg done we will start working again for
induvidual herds. Since sci-physics is done, feel free to send me an email or
talk with me on #gentoo-science so i can add you to sci-physics alias and start
working again.

Thanks

It's time to create sci-mathematics sub project. Should we take sci-calculators
packages too ? Everyone interested please add yourself to bug 139682 and start
discuss about the new sub project.

I'm kinda wondering whether the blas and lapack stuff deserves its own herd.

blas-reference
lapack-reference
cblas-reference
blas-atlas
lapack-atlas
acml
mkl
eselect-blas
eselect-lapack
eselect-cblas
lapack-config
blas-config (both *-config dead soon)

I'd say put them in sci-mathematics, or keep sci-libs.

How about a sci-numerics, or rather a math-numerics, since math-proof is on its
way. With crunching-number and numerical analysis packages:

sci-libs/acml
sci-libs/blas-atlas
sci-libs/blas-reference
sci-libs/cblas-reference
sci-libs/cln
sci-libs/djbfft
sci-libs/fftw
sci-libs/gsl
sci-libs/lapack-atlas
sci-libs/lapack-reference
sci-libs/mkl
sci-libs/scipy

sci-mathematics/glpk
sci-mathematics/octave
sci-mathematics/scilab

From the gentooscience overlay:
sci-libs/arpack
sci-libs/minuit
sci-libs/umfpack

And other that I certainly forget.