Re: Gimp-Print and the implementation of color management
Re: Gimp-Print and the implementation of color management
- Subject: Re: Gimp-Print and the implementation of color management
- From: Robert L Krawitz <email@hidden>
- Date: Tue, 13 May 2003 07:51:45 -0400
Date: Tue, 13 May 2003 14:39:32 +1000
From: Graeme Gill <email@hidden>
As far as I'm aware, noone from the Gimp print/CUPs world has
bothered to spend the couple of hours it would take to add ICC
device link profile support to any of the Open Software printing
systems. [To be fair, Argyll's documentation is severely lacking,
and my choice of a non-standard build system presents an initial
obstacle, but it all seems "to hard" for those few developers who
have started out enthusiastically, and then quickly petered
out. Given that I've done all the really hard work for them, this
is kind of disappointing...] My current conclusion is that there is
a large gap in the determination, financial resources and technical
knowledge between the commercial vendors of color critical
applications, and the "enthusiast amateur open source developers"
that seem to tinker at the edges of various open source
applications. Many of the most significant open source
applications seem to have a core team of programmers who are the
equivalent of any commercial group, and are often backed by some
commercial entity, but generally color is not deemed as a critical
enough feature to get much of their attention.
Integration of third party components is a hard problem, in both the
commercial and the free software worlds. If the documentation's poor
and the build system is lacking, it makes it a lot harder, as you've
seen. I can assure you that this is just as true in the commercial
world as in the free software world. Making it unnecessarily hard to
integrate acts a big barrier to entry.
Part of the problem with Gimp-print is that we've been a bit out of
balance, focusing initially on quality rather than architecture. That
was a conscious decision, and I think it was the right one, but the
downside is that we have some architectural catching up to do in order
to be able to integrate another module. I can certainly see the irony
that this makes it *harder* for us to achieve better quality, but as
they say "the perfect is the enemy of the good". However, now is
clearly the time to take the next step.
My own thoughts are that incorporating various particular CMMs into
Linux/Open Source applications is actually repeating the mistakes
made by commercial operating systems earlier on, and that a better
move in the long run, would be to define an Open Operating System
Color Services module, that would mimic the kind of services
provided by ColorSync, and allow a variety of back end CMMs,
including Argyll, LittleCMS, commercial CMMs etc. (an idea I've
tinkered with, but haven't got serious about yet. That's the
trouble with all this being just a spare time activity :-)
Ultimately I think that's the right way to go. The hard part is
getting there.
There are basically two approaches: top down (devise an overarching
color management architecture, and then push it down into the system),
or bottom up (start with something specific, and generalize). It's my
belief that the latter typically works better in the free
software/open source community. In the absence of an example, nobody
really knows how to design something, and the integration's too much
of a pain for everyone.
My personal take is that the only way things get done from scratch in
the free source world is for someone to do a prototype and actually
make it useful, so that it spreads around a bit. Some of the really
big projects did actually start out with a goal in mind, typically
with sponsorship from a company (e. g. OpenOffice, Mozilla), but both
Linux and Gimp-print started out largely as prototypes, with more
effort going into utility than architecture initially.
In the shorter term, we need to demonstrate just how much difference
color management makes, through a prototype of some kind. Of course,
I'd prefer to have Gimp-print be that prototype, and perhaps we can
figure out how to make that happen.
--
Robert Krawitz <email@hidden>
Tall Clubs International --
http://www.tall.org/ or 1-888-IM-TALL-2
Member of the League for Programming Freedom -- mail email@hidden
Project lead for Gimp Print --
http://gimp-print.sourceforge.net
"Linux doesn't dictate how I work, I dictate how Linux works."
--Eric Crampton
_______________________________________________
colorsync-users mailing list | email@hidden
Help/Unsubscribe/Archives:
http://www.lists.apple.com/mailman/listinfo/colorsync-users
Do not post admin requests to the list. They will be ignored.