Larry, you've made some very good points. Most of the folks on the devel
list have been using LyX for a long time, and it's hard for us to look at it
from the perspective of new users, especially ones who don't even know
LaTeX.
That said, I think it's too late to stop 1.0. The LyXers have been saying
for a long time that all LyX really needed was decent LaTeX import and it
could be released as 1.0. Thanks to the work of a few brilliant programmers
(it's friday!) we've now got LaTeX import which I'll daringly call decent.
(Certainly compared to my few attempts to import things into Word. At least
here everything stays in ASCII!) Given the number of people that I've
personally heard say "I would use it if it were 1.0", and recognizing that I
have that attitude myself, I think releasing as 1.0 is important right now.
Especially since folks have suffered through the pain of a feature freeze
and prereleasing for so long.
I agree that the somewhat limited textclass selection (at least compared to
what can be done) and example/template selection will make LyX 1.0
frustrating to use for some users. (But any college student who wants to
write letters and class papers shouldn't have much trouble.)
Perhaps we need to retarget 1.0. 1.0 may not take over the world. As you
said, we still can't hide ERT. 1.0 should maybe still focus on people who
have used LaTeX (e.g. the scientific societies that have LaTeX classes), and
Linux users, and other people who are comfortable with actually having to
learn a bit about a program. The truth is that while I think 1.0 is great,
and all the devvies deserve praise for it, there's just no way it would get
into mainstream press, and it *shouldn't* yet, because the mainstream press
would all say it's way too hard and noone would try it. Remember, they're
just starting to get used to Linux. ERT would drive them crazy.
Here's what I would suggest. Release 1.0 now. Then, concurrent with the 1.1
development, begin the "formal effort" you describe of adding many new
examples/templates and textclasses, as well as providing up to date docs for
it all.
Asger brilliantly points out that the folks on the devel list are too polite
to fill others' inboxes with "me too" messages. While most of the devvies
are (I think) programming in large part because it's fun and an interesting
and challenging project, all of us are excited to see the LyX user base
growing. And all of us recognize that the user base will grow if we make it
easy for new users to join in. It's just that most of us don't have the time
to organize that.
Which leads to the next problem. Asger suggested uploading your example
files, but---while a noble idea---that probably won't lead to everyone else
uploading their example files. You may remember (or find in the archive) an
extremely brief thread I started called something like "Example files
contest" where I suggested that a contest might be a way to get people to
send in files. The prize, of course, is that your file gets distributed
worldwide. OK, it was a stupid idea. But the point is, the textclass/example
effort will go nowhere without a leader. And of course, you're the logical
leader, Larry. (1) You care a lot about it. (2) You're not (AFAIK) a code
demon. (3) You speak/write English well, which is essential for non-code
stuff. (4) Layout file syntax is simple and the texperts on the list will help
you with cls stuff. And most importantly (5) you're the one who brought up
the problem.
Without active leadership, net projects die quickly. Remember the doc
project? Other than Mike, who (including me!) has written any docs lately?
And yet I could've sworn the idea was to have the docs set up perfectly for
v1.0.
So. I hope you'll take this historic opportunity (sorry: state of the union
address on the brain) to start a textclass/example project. I think there's
a lot of good things that would come out of doing this project concurrently
with 1.1 development. In addition to being the cleaner kernel version,
1.2 could be the World Domination version.
(1) ERT will be hidden in 1.2. Easier to make templates
(2) Toolkit independence means we can interface with KDE (and gnome?), which
means LyX can be just another application you double click on from your
Mac-like desktop.
(3) Working concurrently with 1.1 means that devvies who *can't* really
contribute much to the T/E effort have something to do, instead of shutting
the whole project down.
(4) Working concurrently with 1.1 means that you may be able to make
specific feature suggestions which can be integrated during the earlier
stages of development (e.g., now). Specifically, feature suggestions that
will make things easier for new users. Either LyX functions, or new Layout
tags, or whatever.
The WD release, then, would be the one where we'd pull out all the stops in
the PR campaign. Where we would do our best to get even non-Linux media to
review LyX or publish the press release. And *that* press release can focus
a bit more on completely newbie ideas.
Wow. Looks like this was a long e-mail. But I think you have a really good
idea, and a really important one. Now is the time to make sure LyX
stays/becomes more accessible to newbies. But if noone makes the project
their own, it won't happen.
-Amir
p.s. By the way. It sounds like linux kernel 2.2 will be coming out next week.
Not such bad company, hm?
