Re: Metaproject: Debian Med tasks
On 30.01.2010, at 16:31, Andreas Tille wrote:
> On Sat, Jan 30, 2010 at 01:57:48PM +0100, Holger Schmuhl wrote:
>>>> In general I have quite some ideas about how to enhance my list (if I would have the time beside my 'normal' job ;-) ) ... maybe this could be also interesting for the Debian-Med task list or some joined efforts (?):
>>>
>>> The idea to join forces came to me before I contacted you. ;-)
>>
>> So ... where to start?
>
> That's hard to say. IMHO it would be a good idea to start a common
> discussion on a mailing list. But I think we should not start with
> these general discussion list and it should rather be focussed on
> implementation of code. I remember that I was subscribed to some lists
> where I became quite bored and I stopped subscribing these. I would
> invite all people who are interested in implementing medical software in
> some kind of packages (not necessarily in Debian) to at least subscribe
> to the Debian Med list lurk around how we try to handle problems and be
> not afraid to report about solutions or problems in other distributions.
> The Debian Med packagers should try to focus more technical discussion
> consequently on the Debian Med Packaging list to not be to specific
> for outsider and users.
>
> For the creation of a common list of medical applications I'm absolutely
> sure that a manually maintained list is not maintainable over a longer
> time. We just tried it on static web pages and in a Wiki and I'm afraid
> that our prospective package list (the part of the tasks pages at the
> bottom in red layout) will also contain links to project which will not
> be worth linking any more sooner or later.
Maybe a bit off topic ... but if there is a kind of common list and certain content is pulled dynamically form other sources like the distribution specific listings. Still there is the need to add new projects or to remove dead ones or to update some static content.
How should this be maintained? Everyone has access and can do this (like on Wikipedia) or just a closed group of people/admins? Or by a kind of moderated approach in between, which would mean that everyone can post additions/corrections but they have to be approved by admins.
> So we should rather relay on
> projects we have a grip on via the quality control of a distribution
> (Debian for the moment) and saying this I could easily export a
> structured file (yaml?) with the information of our tasks pages. You
> might either move this to a database or just use this file as the basis
> to generate your own list. The advantage for you is that you can base
> your list on some kind of quality assured part while we might profit
> from some help in the part of not yet packaged software (the prospective
> packages). We could use help maintaining these, picking the projects
> which would be worth packaging with high priority to attract more users
> etc.
Ok, I see your point. Beside the manual export is there also the chance to get a direct integration to the distribution specific database? I mean to show the packaging status, version numbers, screenshots, license types etc. also on non-debian specific web site like mine?
For maintaining not yet packaged software and highlighting interesting projects to be packaged next, should I send an info just to this list in case I stumble upon some interesting stuff?
>
>> I mean a vision could be to have just a single 'list' on the web. Information like project description, license type, links to support-forum/documentation & publications (usage reports), screenshots, activity status etc. are of use to all (Debian, Fedora, OpenSUSE) and can be gathered & maintained in a collaborative effort. Additionally we could have a distro specific 'section' which is used for packaging coordination and stuff like that.
>
> Yes, something like that. As I said above I would offer to export some
> structured information as a basis for this list. I do not see a future
> in a system which is invented *in* *addition* to the work of the
> distributors because this means manual work. I believe in a collection
> of automatically generated information from the info pool of the
> distributors (perhaps by enriching it by a better solution of not yet
> packaged stuff). If we are talking about distributors we should not
> forget FreeBSD which have some interesting stuff as well[1].
Ok, so I hope that I can improve my list some day and evolve it from a static flat list to a dynamically generated list ... does anyone have a suggestion of some kind of useful CMS for especially this purpose? Otherwise I once looked a bit into Ruby on Rails and this could be also a feasible approach to it, although this would mean a bit of work to create it form scratch ...
>
>>> We are currently working on the publication issue to insert this
>>> information into the packages itself in a structured way (I could send
>>> links to the discussion if needed).
>>
>> In case I found some I added it to my list (freemedsw.apfelkraut.org): Publications are labeled by "Userful <book icon>:" ... but I do not have so many.
>>
>> And yes, please send me the link to the discussion.
>
> Perhaps this is a useful start:
>
> http://lists.debian.org/debian-blends/2010/01/msg00000.html
Thanks, I will take a closer look.
>
> (Charless might comment on this.)
>
> Kind regards and thanks for your interest in the work of Debian Med
You are welcome, I really like the idea of making FOSS more popular in the world of medicine & health care and IMHO to offer such applications ready-to-use via a distribution plays the key role in this endeavor.
Best regards,
Holger
Reply to: