>>>>> "Matthias" == Matthias Ettrich <[EMAIL PROTECTED]> writes:
Hello there,
I've been reading this thread when it was nearly over, so all the
technical points have been taken. Since I am not going to do that
again, I'll restrict myself to the subjective part :)
Matthias> Of course. Restricting LyX to one toolkit is not a
Matthias> restriction in terms of software engineering, but a design
Matthias> decision that leads to more flexibility - and thus to faster
Matthias> and more innovative software development.
That would be true if "innovation" (didn't you borrow this term from
B. Gates?) for LyX was mostly themable interface and paperclip wizard.
I hope that the real innovation in LyX will come from the
functionality it provides.
Matthias> What is more
Matthias> restricted, a computer program that uses either C or Pascal,
Matthias> or one that tries to support both languages from the same
Matthias> codebase, with the help of certain macros and some
Matthias> preprocessing?
One could cite TeX as an example of such a restricted (and
non-innovative) program.
Matthias> John, wake up! See how long you guys have been working on
Matthias> that and see how far you came until now.
A long time has been needed to change LyX from a C++ program working
with C library to a C++ program using C++ tools (no offense on the
original program intended). If this is not using modern technology
(yadda, yadda, insert usual buzzwords here), then I do not not know
what you mean. Or maybe you just mean "visibly modern" technology, in
the User Interface.
Matthias> Of course what you
Matthias> describe *can* be done. But it's even more effort in terms
Matthias> of development hours than doing two forks and let one use
Matthias> Gtk and one Qt. In practice, it simply will be a low common
Matthias> denominator. Maybe not the lowest, but pretty low.
Maybe, we'll see. Just remember that the interface is just a _tool_
that the application uses to make the user's life easier. LyX needn't
become a showcase for all the latest eye candy that kde/qt offers.
Matthias> With this argument, we still would avoid using toolkits at
Matthias> all (some users don't want to use XYZ toolkit, so better use
Matthias> Xlib. Ooops, some users don't want to use X, so better use
Matthias> Curses. Oops, some users have broken termcaps and cannot use
Matthias> Curses, better use stdout. ...)
Come on, you know what he means. You have been arguing in another
message that nobody has anything less than a Pentium 300 nowadays, and
that those who do can go to hell. I personally refuse to develop a
program which does not run on my pentium 120 (which runs happily kde
1.1.2, but I begin to wonder whether kde2 will be a good idea).
A side note: you say it is a pity LyX does not let you edit files over
a network. Beside the fact that this is easily fixable, I am ready to
bet that there are much much more people using LyX on an old pentium
than people who want to edit their LyX files over ftp.
Matthias> You cannot argue away the fact that three years ago, the Qt
Matthias> port was stopped due to purely policial reasons (read: Qt
Matthias> was not available under the GNU GPL).
What's the problem with political reasons? You mean we are supposed to
develop software and put no value in it? That's not how I see things.
Anyway, in this particular case, the problem was also that there was a
possibility that Qt would just die from these political reasons, and
that we did not want to sink with the Titanic.
Matthias> This is not about reducing the choice of the user, this is
Matthias> about giving the user the choice to use a modern LyX. Today,
Matthias> the user cannot chose anything, because it simply isn't
Matthias> finished (nor will it be anytime soon).
You seem to forget the fact that people actually use LyX to write
books, and publish them. The goal of LyX is to be _useful_, not
modern. Part of this usefulness is to have a decent interface, of
course, and LyX needs one. But this is not the alpha and omega of LyX
future. KWord has a modern toolkit, a modern interface, devoted
developers, financial support. Is it now to the point where people
actually use it for big documents? I do not mean to insult the KWord
developers, and I do not know much about it besides hearsay, but I
am not sure it has the same maturity as LyX (which would not be
surprising, since LyX had a headstart).
Matthias> I don't have any commercial concerns. But I do have a
Matthias> concerns in using a decent LyX version in the year 2000.
I guess you meant "a version of LyX with a decent interface", right? I
can't believe you are part of the people who judge a program just by
its icons.
Matthias> Believe me or not, I'm thinking in LyX' interest here, not
Matthias> KDE's.
Are you? You say that LyX should become a "killer application". What
is a "killer application"? It is an application that makes the Coca
Cola Company drop windows+office in favor of linux+lyx. Does LyX need
to be a killer application? I do not know, we have to see whether the
quality of the program will improve due to that. The world will not be
a better place because Coca Cola is happy. I will not become richer
either [which could have been a valid reason, of course :)]
What I know is that TrollTech, KDE, Gnome, Redhat, Suse, Caldera and
the others do need killer application. But that's not my problem,
personally.
Matthias> Now, if I get a statement from the LyX team that says
Matthias> something like "we are not interested in providing a decent
Matthias> text processor anytime soon, but we want to invest our spare
Matthias> time in inventing new ways of programming for different
Matthias> toolkits and coming up with new abstractions", that's fine
Matthias> with me.
We are interested in providing the best document processor. Interface
is not what will achieve that. It is part of it. But if we bet all the
future of LyX on the interface (the "sexy" part), we will spend the
next few years dealing with the complaints of people who cannot embed
MP3's in LyX and have them play on the printer...
Matthias> Seriously, in the free software business, you cannot make
Matthias> everybody happy. There will always be people complaining.
Sure. That's why those who write the program get to make the decisions.
Matthias> For your best sake, you have to make decisions. Programming
Matthias> language, scope, toolkit, platforms are some of those.
Sure. That's what we do. Just that these are not the decisions you
would have taken.
Matthias> Keep in mind: a complete KDE port that replaces all the
Matthias> xforms stuff with Qt/KDE code can be done in two weeks work
Matthias> if you have at least two people fulltime. With the
Matthias> abstraction layers in there, it might be more complicated
Matthias> now than it was with LyX1, but one could simply drop some of
Matthias> those and replace them with the ones Qt already provides.
You mean that we would have the same spaghetti code than before, but
with Qt instead of xforms? I'm not sure the gain is so high. In case
you have not remarked, most of the work done for LyX 1.1.x is internal
rewrites. It does not give you PC Magazine Editor's Choice Award, but
it is needed sometimes.
Matthias> It would
Matthias> simply be the standard way to write letters and longer
Matthias> reports on Linux. And what do we have now? A great tool for
Matthias> those scientists that are not scared to death by XForms.
As one of these damn scientists, I tend to think that people who are
scared to death by xforms are a bit easily impressed. We are not
talking about using 'vi' here. Xforms has an interface that is
certainly ugly, but mostly as functional as Qt. I did not see many
things besides the math toolbar in KLyX which really improved user
interaction. Sure, LyX with Qt widgets is better that LyX with xforms
widgets, but I have not been impressed by the KDE integration.
Matthias> Think about your users and what you really want to develop,
Matthias> not about politics.
A few years ago, many of those users chimed on the mailing list
saying: "why don't you just rewrite LyX in Java? It is a modern
language and all of you portability concerns will be gone". Should we
have taken this route? I doubt it. This was however a clear case of
modern technology that users want. What I want to say here is that
before writing a sentence like yours above, you decide what is good
for you users: have a program which uses all the latest Qt 2.2.x
widgets.
I do also have my own conception of what is good for our users, I I do
try to force it into their throats: I try to force them to give a
structure to their documents, think in terms of what they want to
write instead of what it looks like; in short, use LaTeX instead of
Word. I think this is a much worthier goal. And yes, this is a kind of
political goal; I see LyX as a an educational trojan horse: make
people learn how to write their documents when they are just trying to
know what icon will give them bold for this damn section heading.
Matthias> As a nifty side effect, a Qt port also offers the
Matthias> possibility for a native MS-Windows version (soon also
Matthias> Macintosh).
Is it native as in "native windows widgets"? I thought Qt handled all
the drawing by itself (which is less desirable from a "native" point
of view).
Matthias> If some of you want to work on that, I'm sure my
Matthias> company can provide you with the necessary number of Qt
Matthias> licenses for MS-Windows.
And then how will people be able to use it? If we plan to continue
supporting windows, I'd like to be sure that it will be as free (in
the RMS sense) as using MFC, for example. Remember that half of the
reason we want to go away from xforms is that it is closed source and
has endless build/versions problems.
Have a nice week. I am glad to report that you failed to ruin my week
in Venice, although I spent way too much time thinking about it...
JMarc
PS: I understand that many things that I wrote are either wrong, of
dubious interest, insulting or stupid. However, your messages on these
subject have the curious effect that they give me extremist ideas I
would not have had before. Today I think "phew, I'm glad we did not give
up, these guys are dangerous!" :) Over time, I will probably
come back to more reasonable thoughts. Until your next message.
PPS: If you really want to convince me/us, you should probably re-read
Machiavel. With the wording of your message, I nearly thought you took
some ideas of B. Gates on Linux and changed the words to make a good
prank. Unfortunately, I was wrong.
