Executive summary: What you and Mats and everybody who thinks a cross-
platform toolkit is an attractive plan need to do to get me on board
with that idea is this. Start evaluating the existing toolkits. Don't
worry about the technical lowlevel stuff for integration. Which ones
look attractive on Linux/*BSD (KDE & GNOME desktops), Solaris,
Windows, and Mac? Which ones provide functions to pop up printer and
font-selection dialogs on all supported platforms? Etc. You know
what you think is needed for desktop integration, so you can at least
weed out the ones that don't even satisfy your basic requirements.
Give me that much, and I'll put my experience and knowledge behind
your idea, instead of resisting it. But I need some direction,
because I don't think it's worth it, and as it stands I'm not willing
to put out the effort to specify requirements, investigate features,
and evaluate technical feasibility. Give me a leading candidate (on
the grounds of attractiveness and providing the relevant high-level
features. Then I can concentrate on that last task. If it at least
comes close on technical grounds, we can go on to second choices in
hopes that they're technically sufficient.
I don't speak for the Board on this; that's a personal commitment.
But if you can satisfy me, you'll have a pretty strong advocate on the
Some additional comments:
Marcus Harnisch writes:
And that's exactly why we need to outsource stuff.
If this were proprietary software, "what we need" would be backed up
with money, and money would reasonably quickly elicit what we need
from the market. It's not, so "what we need" and "what we will
have very little relationship to each other. What I've seen "out
there" leads me to believe there's not much we can outsource that we
haven't already done.
> GUI users don't become Emacs developers.
I disagree. Every developer starts out as user.
True, but so what? We're not talking about the conversion rate for
all users here. We're talking about the conversion rate for users
attracted to XEmacs by desktop and platform integration. I don't know
of any such developers (and that's true for Emacs too), though there
may be some. 0% times 1,000,000 users is still 0 developers.
> And IMO we should not rely on toolkit functions. They are not
> sufficient to do what we need for proper Unicode support in MULE,
> while still supplying faces and all that stuff. What we need is
> something on the level of the Mozilla redisplay, "gecko", not the kind
> of simple widgets provided by toolkits.
And Mozilla uses toolkits for widgets, no?
Yes. Exactly! Multiple toolkits. Each platform's toolkit. Just
like <effect type="drum roll">XEmacs does</effect>, and you
we should stop doing. :-(
> I think there's a common misunderstanding of what a toolkit
> actually provide for applications like Emacsen and web browsers.
Look, Marcus, I've read all of the redisplay code. I've dug deep into
the Xt event loop. I've worked on Xft. I've worked on Mule. I was
there when Jareth Hein spent 3 months replacing all our image code
with libmagick, and then the next three weeks ripping it all back out
and replacing it with (improved) lowlevel marshalling code on a
library-by-library basis because libmagick sucked as a way to handle
images for XEmacs display. I understand the theory, but I've also
participated in the practice.
I accept the theory as correct, as far as it goes. I acknowledge that
my experience is hardly complete. But since it's my time that's being
demanded here, I'm going to insist that my negative experience with
"cross-platform" toolkits trumps that attractive theory. "It's a good
idea" is not enough to get me to move; we need a concrete proposal.
And I don't have one to offer myself.
That same developer would spend his/her time better with
implementing *an additional* feature rather than one feature for a
number of platforms.
Olivier already answered that. In the opinion of the people who have
done this kind of work in the past, and presumably will be called on
to implement this proposal, dependence on a cross-platform toolkit will
dramatically decrease the quality of the features we can provide.
XEmacs-Beta mailing list