Re: Metaproject: Debian Med tasks
[now more longish answer after short notice of problem with tasks pages]
On Mon, Feb 08, 2010 at 11:07:07AM +0100, Sebastian Hilbert wrote:
> 1.) software reagarding medical/scientific tasks are spread all over the web.
> Effort has been undertaken to aggerate those in one place. Personally I have
> lost track. There are projects at freshmeat, sourceforge, savannah, debian-med
> and whatnot.
>
> I personally feel it is up to the project maintainers to make their work
> public. Sure it would help if there was this one place where all software
> could come together but this is the missing link so far
I agree here. That's why I would not try to make an effort to "just
another link collection" which are in principle a form of advertising.
Our aim should be the "one click to install and run" ... which is a
marketing clause, but in principle we try to shorten the way from
learning about a project (by reading a description on something like our
tasks pages) to fire it up and run it. IMHO this only works in a Debian
Med like (it must not be Debian, but a solid and reliable distribution
with a forseable future) way. If there is the chance to synchronise
packaging efforts between distributions everybody might profit.
> 2.) I did not find GNUmed listed at Debian-med any more. Did I just overlook
> it or miss something ?
Explained in previous mail. Nearly all packages are currently missing in
UDD. I suspect some strange disk failure (ftpmaster was reporting vaguely
about a broken patrition on some Debian infrastructure machine - no idea
if this is connected).
> 3.) There seems to be an unofficial Debian package of openemr you might want
> to take a look at
>
> http://www.openmedsoftware.org/wiki/OpenEMR_Downloads
That's a good hint. Most probably we should step in here and try to make
this official.
> Summary: Visibility of the great software and packaging effort of FOSS
> software is far from ideal. This severly hinders your chances to get picked up
> by users.
>
> Everyone and their uncle has an app store today. All but FOSS software. Which
> is a shame. I mean first we create an kick ass alternative software. Then we
> either do not package it at all and require users to have Jedi powers to get
> it running or we package it in far too many formats and distributions so that
> a single prospective user does not have the slightest chance to get it
> installed. Why the hell do I need deb (Debian/Ubuntu etc.) rpm (individual rpm
> indeed for openSUSE, Fedora, Mandriva etc.) ebuild and whatnot ?
Common, Debian Med is one way how Debian will gain world domination so
this part of your problem will be solved soon. ;-)
Honestly: In how far do you think that freedom makes your live *easier*?
So people take the freedom to make their own distribution. There is no
other chance than *proving* that one distribution is superior to all
others and I do not think that this will be possible at all. So there
is some diversification which makes life harder, but there is no handle
to push people in one direction. That's just life and life is not that
simple.
> Come on I am a software developer. Currently I spend more time packaging
> software then developing. This got to change.
I for myself try to do my best to take the burden of Debian packages
from your shoulders. That's my/Debian Med's share in this business.
After beeing a Debian Developer for > 10 years I really wonder whether
it is really possible to do much more. The thing is that there are much
specific things to respect and that you need to know a lot of things to
make reliable packages. So I think we need a *cooperation* between
upstream developers and packagers not a single person who does both jobs
alone. The additional positive side effect is that more eyes are
locking onto the code which turned out to be positive in the past (even
if my contribution to GNUmed problems was low, it happened to some
extend).
> I know this is good for a flame
> war. Please don't. Let us find ways or maybe an infrastructure where packagers
> from various distributions can collaborate.
I hope that this will happen, yes. But I doubt that it will be the
final solution of your problem as I tried to explain above.
> Tell you what. The minute the Linux/FOSS community decides on one package
> format we will see adoption rate go up. Distributions could concentrate on
> innovation rather then packaging and software developers would be relieved of
> the packaging job and hunting around for the right place to send the software.
While I agree in principle I doubt that this will practically happen.
BTW, it is not only the package *format* RedHat rpms are different than
SuSE rpms (as far as I know) and one important critics which frequently
leads to flamewars is that Debian and Ubuntu debs are not necessarily
binary compatible. So it is not the format, it is way more and we will
not solve this problem on Debian Med list. The only thing what we can
do *here* is to make Debian in itself good enough that people working
in the field of medicine do not start just another medical distribution.
For instance I have a small excercise for you. Just go to
http://nebc.nox.ac.uk/tools/bio-linux/bio-linux-5.0
and have a look at BioLinux. Previous versions were derived from Debian
now it is derived from Ubuntu. I actually do not care from what source
these people derive - I care that they derive at all and do not try to
plug in there work into a distribution. I've wrote some emails in the
past, I'm even advertising their (non policy conformat Debian packages
as unofficial packages on our tasks pages, each with a dedicated hint to
BioLinux). My assumption was that we might be able to cooperate and an
upstream developer who has his office next to the developers of BioLinux
perfectly agrees with me that a cooperation makes sense. Guess whether
I've got any response?
So my way to go is: We must be cover *all* of their software (OK, there
is some non-free binary stuff we can't) and make sure it runs more
reliable, which meens we need to *prove* that we are better. That's a
hard job because these people are not actually bad, they just disagree
that taking over Debian Med work and sticking to Debian policy is a
reasonable thing to do.
Now try your e-mail upon these people and try whether your approach to
convince people by words instead of patiently doing good work as a
little sandbox experiment of the larger unification you might have in
mind.
Kind regards
Andreas.
--
http://fam-tille.de
Reply to: