I must admit thats exactly what I had in mind.
Joschi has put me off a bit with his estimate of how much work it will be....
mark
-----Original Message----- From: Rory Walsh [mailto:rorytheroar@yahoo.com] Sent: 16 April 2002 13:51 To: pd-list@iem.kug.ac.at Subject: Re: [PD] Call for Unified Libraries
Actually would a better idea be to leave the large libraries such as zexy as they are and work on a library that puts together all the other little externals from other people who are just beginning to develop objects for pd? Would this make more sense?
--- Joschi joschi@eds.org wrote: > This is a good idea, anything to ease the
installation of a good PD system for the end user is great. I do think that the challenge of this is that it is going to take a committed team to keep up with all the parrallel development that goes on in the pd community over a long period of time. There is no point in starting such a project unless you are willing to spend 5-10 hours a week for the next 2 years at least. There are many people working on externals in their own bubble, at their own schedule, and I don't think that is going to change, so what it is going to take is a committed group of people who want to invest serious time into bundling up all this exciting work into an easy to install form.
On Mon, 2002-04-15 at 05:45, mark wrote:
Hi,
I have just rebuilt my laptop and the biggest
problem
with PD is gathering together all of the externals
and
documentation.
It occurs to me that unified libraries would be
the answer -
probably hosted on sourceforge and gathering
together
as many externals as possible.
I guess there would have to be a number:
non~tilde (nt) cross platform tilde (t) cross platform
nt - linux only t - linux only
nt - win only t - win only
How does this strike people? Would all you
external writers
contribute?
cheers
mark
Do You Yahoo!? Everything you'll ever need on one web page from News and Sport to Email and Music Charts http://uk.my.yahoo.com
I really dont intend to put anyone off doing this! That was just a quick dashed out estimate, really the main point i am trying to make is that the _key_ to a successful free software project is stability, all the successful projects (gnu, pd, gnome, apache) had core contributers who stuck with it through the long periods before others started to come along and chip in. I think some people come into a free software projects and see all the help from outsiders and think "hey, if i start a project, others will come in and help with the leg work" but that isnt the case 90% of the time.
This project should happen and I think that if a team is put together of people who are willing to commit a certain amount of time to it for the long run, it can succeed.
:)
On Tue, 16 Apr 2002, mark wrote:
I must admit thats exactly what I had in mind.
Joschi has put me off a bit with his estimate of how much work it will be....
mark
-----Original Message----- From: Rory Walsh [mailto:rorytheroar@yahoo.com] Sent: 16 April 2002 13:51 To: pd-list@iem.kug.ac.at Subject: Re: [PD] Call for Unified Libraries
Actually would a better idea be to leave the large libraries such as zexy as they are and work on a library that puts together all the other little externals from other people who are just beginning to develop objects for pd? Would this make more sense?
--- Joschi joschi@eds.org wrote: > This is a good idea, anything to ease the
installation of a good PD system for the end user is great. I do think that the challenge of this is that it is going to take a committed team to keep up with all the parrallel development that goes on in the pd community over a long period of time. There is no point in starting such a project unless you are willing to spend 5-10 hours a week for the next 2 years at least. There are many people working on externals in their own bubble, at their own schedule, and I don't think that is going to change, so what it is going to take is a committed group of people who want to invest serious time into bundling up all this exciting work into an easy to install form.
On Mon, 2002-04-15 at 05:45, mark wrote:
Hi,
I have just rebuilt my laptop and the biggest
problem
with PD is gathering together all of the externals
and
documentation.
It occurs to me that unified libraries would be
the answer -
probably hosted on sourceforge and gathering
together
as many externals as possible.
I guess there would have to be a number:
non~tilde (nt) cross platform tilde (t) cross platform
nt - linux only t - linux only
nt - win only t - win only
How does this strike people? Would all you
external writers
contribute?
cheers
mark
Do You Yahoo!? Everything you'll ever need on one web page from News and Sport to Email and Music Charts http://uk.my.yahoo.com
Josh .. Yoshi .. Joschi .. xiphoidprocess.com .. eds.org/~joschi
I just felt the need to give forth my 2 cents on this topic:
I don't think that a "Unified Library" is the way to go.
Rather, PD and all of its libraries need to be placed in debian or RPM packages... easy installation, management, dependency resolution, etc will then be possible.
With a couple simple scripts and a howto we could make it very easy for authors to package up their externals.
The Demudi folks would probably be happy to include this stuff on their server/in the distro.
Unfortunately this won't help the Windows folk... All I can say on that is that life sucks when you are locked into a proprietary system... may as well compile a giant DLL and pray that everything works. *grin*
If someone has a lot of time on their hands, a package management system could be integrated directly into PD.
andy
i agree that it just makes more sense to keep each library in a separate .dll... in the windows world something like the cygwin installer could be snaked to deal with managing each package...
On 16 Apr 2002, Andrew (Andy) W. Schmeder wrote:
I just felt the need to give forth my 2 cents on this topic:
I don't think that a "Unified Library" is the way to go.
Rather, PD and all of its libraries need to be placed in debian or RPM packages... easy installation, management, dependency resolution, etc will then be possible.
With a couple simple scripts and a howto we could make it very easy for authors to package up their externals.
The Demudi folks would probably be happy to include this stuff on their server/in the distro.
Unfortunately this won't help the Windows folk... All I can say on that is that life sucks when you are locked into a proprietary system... may as well compile a giant DLL and pray that everything works. *grin*
If someone has a lot of time on their hands, a package management system could be integrated directly into PD.
andy
Josh .. Yoshi .. Joschi .. xiphoidprocess.com .. eds.org/~joschi
Andrew (Andy) W. Schmeder hat gesagt: // Andrew (Andy) W. Schmeder wrote:
The Demudi folks would probably be happy to include this stuff on their server/in the distro.
The Demudi folks already have! I have a debian package here called "pd-externals" and put together by Guenter Geiger, which includes a whole lot of externals: ggee, ext13, zexy, plugin~, ann,... But probably not in the newest versions, and that would be a problem for any packager: Keeping track of new versions of libraries. I think, a pure-data.org like approach is in that case a better solution, but maybe we need some collection place where developers could announce and maybe upload new versions of their externals by themselves. Like here on this mailing list: I think all externals are announced here. So maybe if we could agree on a new tag in the subject line instead of [announce] for external announcements like [EXTERNAL], one could automatically create a list of new externals from the mailing list archives.
ciao,
__ __
Frank Barknecht ____ ______ ____ __ trip\ \ / /wire ______
/ __// __ /__/ __// // __ \ / / __ \ ___\
/ / / ____/ / / / // ____// /\ \ ___\____ \
/_/ /_____/ /_/ /_//_____// / \ \_____\_____
/_/ _\
The Demudi folks already have! I have a debian package here called "pd-externals" and put together by Guenter Geiger, which includes a whole lot of externals: ggee, ext13, zexy, plugin~, ann,... But probably not in the newest versions, and that would be a problem for any packager:
Yeah.. I know... I've got those .debs. The problem is that they are very out of date (pd 0.33)... also there is no pd-dev package w/ appropriate headers... its just not enough to maintain a decent pd install purely within the confines of the apt- system.
Furthermore pd-externals contains a whole slew of different libraries, which is bad from the maintenance point of view, IMHO.
A good example is the xmms packages. There is no 'xmms-plugins' package, rather each plugin has its own package. Very clean that way... instead of pd-externals we should have pd-zexy, pd-ggee, pd-gem, pd-iemlib, etc. Of course not everything will be packaged nicely but the major libs could be.
Maybe when I've got time I'll stop advocating and start packaging ... but not now. ;)
think all externals are announced here. So maybe if we could agree on a new tag in the subject line instead of [announce] for external announcements like [EXTERNAL], one could automatically create a list of new externals from the mailing list archives.
Now that is a great idea!
andy
Ok, these pd-externals were put together in a hurry (as almost everything regarding demudi), and it is in fact not the way to keep the things managable (there has to be a new package for every lib that has a new version ..).
It seems that the number of contributors has reached a critical mass, where we really have to do something in order to keep the "mess" that is created within bounds.
IMO the ideal solution would be to build a CVS server for externals. Together with some rules how to contribute.
A .tar.gz and zip snapshot from this can then be used for distribution, binary versions for the different platforms can be build from that.
Well, setting up such a system is a bit of work, and it can only succeed if the external developers agree to use it.
This way we could somehow try to start to have a quality control for externals, bug tracking system and the like. This should do away with the concerns Yves has about the pure-data.org site.
Lets not bitch at each other, its not jfm3's fault that external developers are as chaotic as (some of us) are.
In the current situation it is very hard to keep up with the things that are going on, jfm3 is putting a lot of his time into it, with a common effort everybody should benefit, and above all the music we want to create. (... arghh, that sounds pathetic ..)
Guenter
On 17 Apr 2002, Andrew (Andy) W. Schmeder wrote:
The Demudi folks already have! I have a debian package here called "pd-externals" and put together by Guenter Geiger, which includes a whole lot of externals: ggee, ext13, zexy, plugin~, ann,... But probably not in the newest versions, and that would be a problem for any packager:
Yeah.. I know... I've got those .debs. The problem is that they are very out of date (pd 0.33)... also there is no pd-dev package w/ appropriate headers... its just not enough to maintain a decent pd install purely within the confines of the apt- system.
Furthermore pd-externals contains a whole slew of different libraries, which is bad from the maintenance point of view, IMHO.
A good example is the xmms packages. There is no 'xmms-plugins' package, rather each plugin has its own package. Very clean that way... instead of pd-externals we should have pd-zexy, pd-ggee, pd-gem, pd-iemlib, etc. Of course not everything will be packaged nicely but the major libs could be.
Maybe when I've got time I'll stop advocating and start packaging ... but not now. ;)
think all externals are announced here. So maybe if we could agree on a new tag in the subject line instead of [announce] for external announcements like [EXTERNAL], one could automatically create a list of new externals from the mailing list archives.
Now that is a great idea!
andy
IMO the ideal solution would be to build a CVS server for externals. Together with some rules how to contribute.
A .tar.gz and zip snapshot from this can then be used for distribution, binary versions for the different platforms can be build from that.
Well, setting up such a system is a bit of work, and it can only succeed if the external developers agree to use it.
This way we could somehow try to start to have a quality control for externals, bug tracking system and the like. This should do away with the concerns Yves has about the pure-data.org site.
Lets not bitch at each other, its not jfm3's fault that external developers are as chaotic as (some of us) are.
I think it's just the same - contributing to a CVS server or submitting links to pure-data.org, although the latter is sure easier for people with less programming experience. And if a source management system, it might be most convenient to take an existing one, like sourceforge.
As i understood it, there is a possibility to outsource the task of checking the submissions to pure-data.org, so that the work load is distributed on more than just jfm3 and response times can be shorter. I personally like the pure-data.org system.
greetings, Thomas
On Wed, 17 Apr 2002, Thomas Grill wrote:
I think it's just the same - contributing to a CVS server or submitting links to pure-data.org, although the latter is sure easier for people with less programming experience.
Can you explain in what way it is the same ?
And if a source management system, it might be most convenient to take an existing one, like sourceforge.
Well, sourceforge is perfectly ok, did I say something different ?
Guenter
As i understood it, there is a possibility to outsource the task of checking the submissions to pure-data.org, so that the work load is distributed on more than just jfm3 and response times can be shorter. I personally like the pure-data.org system.
greetings, Thomas
On Wed, 17 Apr 2002, Thomas Grill wrote:
I think it's just the same - contributing to a CVS server or submitting links to pure-data.org, although the latter is sure easier for
people with
less programming experience.
Can you explain in what way it is the same ?
You have to submit your stuff.
And if a source management system, it might be most convenient
to take an
existing one, like sourceforge.
Well, sourceforge is perfectly ok, did I say something different ?
I once made the same suggestion and it was turned down (not by you).
greetings, Thomas
-
Andrew (Andy) W. Schmeder -
Frank Barknecht -
g�nter geiger -
Josh Steiner -
mark -
Thomas Grill