Re: Chemistry wiki page
On Tue, 7 Oct 2008, Chris Walker wrote:
No. I think they should be included in the task package, but I'm a
physicist and don't want to force my opinion on chemists.
Ah, OK.
In addition to the task packages, I'd like to see a categorised list
something like https://help.ubuntu.com/community/UbuntuScience where
packages are predominantly not repeated.
I have nothing against such a list, but what is the practical use of
it (besides stealing time of people when disputing to which category
a package belongs)?
I do wonder whether there might be some merit in splitting the support
packages off[1] - the physics package contains 5 Matlab/IDL like packages[2], 3
symbolic maths packages[3], 2 data acquisition package and 1 gui package
for fitting data - and clearly these have use in things like engineering.
Splitting is an interesting question I have thought about several times.
All the times I came to the conclusion that it is purely academical
(scientists are addicted to classification tasks - so lets classify!!)
but does not really help bringing the software to our users. You can
not expect a user to immediately understand your classification system
(you will even hardly find agreement here on this list about classification).
The good thing we have done with the tasks files is to extract a set
of 10-50 packages out of a large pool of 20000. Now picking 5 out
of 20 as you are suggesting is overdesign with the pure purpose of
designing something, IMHO.
I don't think we should include every possible package in the task
packages - I think we should only include those useful to a scientist
doing research. If packages are obsolete or poor quality, we probably
shouldn't include them in the tasks. Even better would be a review of
the alternative packages that seem to do the same thing, but that sort
of thing takes a lot of expertise to write.
Yes. I agree we should establish some quality ensurance. There is no
point in propagating bad quality software. But it is you who has to decide
in your field. There is the "Ignore" field in the tasks files which
can be used for packages you would not like to see in the meta package
or listed at the tasks pages. Just add a comment via "Why" to this
entry and we have a record about the issue.
[1] Perhaps the viewing package is the right place for this.
[2] Should I add Yorick to these, or does it fit better with gnuplot,gri ?
[3] maxima,sympy and axiom. I've just added axiom, but have noticed it
on the RC bug list, so maybe I'd be better to remove it.
These are all good questions. I learned in the past is that answering
these yourself and implementing this answer in the code (task file) is
the best way to find a solution. If somebody disagrees strong enough
he will undo or increase your change.
Kind regards
Andreas.
--
http://fam-tille.de
Reply to: