Re: [Gimp-developer] plug-in distribution choices

2001-05-21 Thread Branko Collin

On 20 May 2001, at 16:36, Sven Neumann wrote:
 Branko Collin [EMAIL PROTECTED] writes:
 
  My suggestion is that the following plug-ins belong to the core
  distribution:

[4 rules of inclusion snipped]
 
  Can such a distinction be made?
 
 You are right that we need to make such a distinction, but I don't
 think the rules you suggested make much sense. 

Ouch. :-) I was writing a lengthy reply here, but I realised that I 
could probably sum it up as follows. When deciding what to include 
and what to leave out, you probably should also keep some vague, non-
objective criteria in mind. Actually, defining descriptive and 
objective criteria might prove impossible. What I am arguing for is 
to include a few plug-ins that may not seem to have much value for 
the average user, but that help set apart GIMP from its competition, 
help inspire plug-in writers and that are just fun to have. A good 
example would be Gee-Zoom.

 On the other hand I
 think we should first discuss how the gimp packaging should look like
 in the future instead of tackling the problem which plug-ins go into
 which package.

OK.

 I'll try to summarize some of the ideas that have come up during
 earlier discussions:
[snip what needs to be done]

Can't we use the Debian tool?

Also, it seems to me that this whole things depends on finding enough 
maintainers. Judging by the state of the web site (parts of it have 
not been kept up-to-date for a while now) that is going to be the 
real hard thing to do.

Finally, you may not be able to get around distributing binaries too. 
Windows users are not used to making their software packages, they 
want tools that install the libraries for them ready to run.

Maybe it would be easier for now to rely on the users to be able to 
collect the right packages. 

-- 
branko collin
[EMAIL PROTECTED]
___
Gimp-developer mailing list
[EMAIL PROTECTED]
http://lists.xcf.berkeley.edu/mailman/listinfo/gimp-developer



web site (was: Re: [Gimp-developer] plug-in distribution choices)

2001-05-21 Thread Branko Collin

On 21 May 2001, at 14:58, Sven Neumann wrote:
 Branko Collin [EMAIL PROTECTED] writes:

  Also, it seems to me that this whole things depends on finding
  enough maintainers. Judging by the state of the web site (parts of
  it have not been kept up-to-date for a while now) that is going to
  be the real hard thing to do.
 
 I always wondered why noone stands up and volunteers to take over the
 website maintainance. But it looks as if everyone gets bored doing web
 stuff really soon...

Erm, I would be willing to help some with the web site. It would seem 
a waste not to do what I am best at. Who should I turn to for this?

-- 
branko collin
[EMAIL PROTECTED]
___
Gimp-developer mailing list
[EMAIL PROTECTED]
http://lists.xcf.berkeley.edu/mailman/listinfo/gimp-developer



Re: [Gimp-developer] plug-in distribution choices

2001-05-20 Thread Sven Neumann

Hi,

Branko Collin [EMAIL PROTECTED] writes:

 Am I to understand that there is no recorded instance of this 
 discussion? Well, let's start now, then, so that next time we can 
 point to the mailing list archives.

The mailing list archive definitely has at least parts of this 
discussion. Since we did not come to a conclusion last time, the 
archive is however not very helpful and you are right in pointing 
out that we should start it all over now. 

 First of all a definition of the problem area: what are considered 
 plug-ins? Everything that goes into the directory 'plug-ins'? 
 Anything else?

Scripts are definitely in the same category. In the future we will
also see pluggable tools and its foreseeable that we will face the
same problems there at one point.

 My suggestion is that the following plug-ins belong to the core 
 distribution:
 
 - those that perform a task that the GIMP should have provided for 
 itself or will provide for in the future;

Our goal is to have as much functionality as possible in plug-ins.
This reduces the size of the core, makes it cleaner and improves
maintainability of the core. This makes this rule questionable
especially since a task that the Gimp should provide is much too
vague.

 - those that will help other plug-in authors better understand how to 
 write plug-ins;

We have a plug-in template for this task and I don't see what would 
stop plug-in authors to have a look into the plug-in packages that 
are distributed seperate from the core gimp package.

 - those that will make the GIMP look good when compared to other 
 raster image editors

Only because another program has a certain functionality does not 
make it necessary to distribute the same functionality with core Gimp.

 - those that perform a task the best in its field.

Not a very useful definition either. Those plug-ins will most likely 
be pretty large and only useful to a subset of users. Thus they belong
into a special package.

 Can such a distinction be made?

You are right that we need to make such a distinction, but I don't
think the rules you suggested make much sense. On the other hand I
think we should first discuss how the gimp packaging should look
like in the future instead of tackling the problem which plug-ins 
go into which package.

I'll try to summarize some of the ideas that have come up during
earlier discussions:

A very important point for distributing a plug-in is to have a 
maintainer that feels responsible for it. The core Gimp maintainers
would like to only have a set of basic image manipulation tasks
in the core package and they would feel responsible for keeping them 
up to date, handling bug-reports and discussing patches. Such basic 
plug-ins would include for example Blur, Gauss-Blur, Sharpen, Rotate,
and a set of important file-plug-ins to give only a few examples.

Other plug-ins could be grouped into smaller packages for special
purposes. For example there could be a gimp-web-plug-ins package
that adds functionality like GIF-Save, Anim-Optimize, ImageMap and
others. Such a package would have a maintainer who is responsible
to handle bug-reports and keeping the package in sync with the core.

Installing gimp would then be a process of installing gimp-core
and a set of plug-in packages that fit the needs of the user. Such
an approach has some obvious advantages but also a bunch of 
drawbacks:

The packages would have interdependencies. The web-plug-ins package
might include a gimp-perl script so it would require gimp-perl. Of
course everything in the web-plug-ins package but this script would
work w/o gimp-perl so actually gimp-perl is not required but only
recommened. I can however only think of one binary package system 
that can handle those kind of weak interdependencies (debian).

The packages should not overlap and they should share the menus
intelligently. This would require some interaction between package
maintainers. I think however that this should be doable.

On multi-user systems the administrator who installs gimp can not
decide what packages the users might want. One solution is to 
install them all. This would however lead to overcrowded menus.
A problem that could be solved by a plug-in manager that knows 
about all plug-ins that are available and lets the user choose
what plug-ins she wants to see in the menus.

Having gimp split up into dozens of packages will make installation
difficult. Again a plug-in manager that knows about all available
plug-ins out there, collects and installs them would help.

It turns out we need a plug-in manager. What functionality should
such a beast have? Here's a list of things that came to my mind:

It should read a number of plug-in lists: a list of all available
plug-ins that is distributed with the core gimp and can be updated 
through the web. A list of plug-ins that are installed locally do
not necessarily appear in the menus.

It should have meta-packages of plug-ins similar to the task-lists
Debian has so a user