Ok, too much mail on this thread, I'll try to summarize my answers:
- The only reason for me to stay on this project is that I want to finish
something that I started. In my view, tomcat 3.3 ( or what will be in the
main branch of cvs in about a month or 2 ) will be the "right" thing
based on the current architecture. Tomcat3.2 is better than 3.1, but it's
not yet there.
- Jon, Craig claim that you'll be better served by 3.2.x - well, look at
the code. It's the same as saying that users would have been better if
only bug fixes were done in 3.1.x - and no real development on 3.2. You
can compare 3.1 and 3.2 and decide for yourself. ( and BTW, most of the
development on 3.2 was done by very few people, the same that are
developing 3.3 ).
- Tomcat 3.1, 3.2, 3.3 all share the same _design_. Tomcat 4 has a
different _design_. The code changed quite a bit in 3.x - that's
true. Refactoring is _good_, and improves the quality of the code, but
it's not making 3.3 so different than 3.2. BTW, the analogy with Apache2.0
is wrong - Apache2.0 started from Apache1.3.x code, and it has the same
architecture ( and builds on top of 1.3.x ). And Apache is built on top of
NCSA code - rewritten few times, but there are still traces and ideas from
the original code to show you the evolution.
- There are 2 things that are important in 3.x - one is the fact that it
is the sum of contributions from many smart developers, like
Gal, Glenn, Nacho, Alex, Henri, Larry, Dan ( not to mention James, who
did the initial design). Finishing their work is a form of respect. And
most open source projects succeed by respecting their developers.
Telling that "tomcat3.3 shouldn't happen because we decided we want 4.0 in
and 3.x out" is not a form of respect for people's contributions. The code
that is part of 3.x is result of many developers work, and I don't think
you have a right to just throw it away.
- The other thing I like about 3.x is the design. It's a small core, and
you can build on top of it. And the modules can be developed independently
- so after 3.3 is done, I can implement the 2.3 spec, do performance
improvements, add all the features that 4.x will have - and that without
having to go through all the political mess that is this list.
I plan to do that - and for that I'll continue to maintain 3.3
core ( not that it'll be a big issue - now 90% of the code is part of
modules that can be easily replaced in case some bugs are found ).
- Work on the main branch of tomcat3.x - there are people who seem to like
it, people who see it as a threat. Well, if tomcat4.0 is better - you have
nothing to worry about. If it can attract developers - you have nothing to
worry about. If it can reuse code and features from 3.x in the same way
tomcat 3.3 can ( and will ) reuse code and features from 4.0 - you
shouldn't worry.
- It seems a lot of people are concerned that because of 3.3, 4.0 is not
getting enough developers ( besides those paid to do that).
I must say, it's a valid concern.
Getting people to contribute is a hard job, and you'll not get too far
with arrogance and brute force. Sometimes it's important to show respect to
developers and their contributions.
But what I don't understand is your approach on resolving that. You may
succeed in stopping tomcat3.3 and forcing it out of apache, but I don't
think that would guarantee that people who worked on 3.3 will just move to
4.0 like little lemmings.
- Probably the thing I dislike the most on this project - the decisions
are made by political means, by few people who have bigger voices or
are better at "lobbying". Is this going to attract developers ? I still
think that code matters and good code can't be destroyed by politics.
- In any case, I'll try as much as I can to finish what I started.When
it's ready I'll use my right to propose it as 3.3. After that I'll try to
have as little as possible to do with apache.
- Many thanks to those of you who sent me private mail. It helps me a
lot. Many thanks to those who sent mails publicly.
- And again, if you are interested in creating a high quality servlet
container, based on the work of many developers, well integrated with
Apache, IIS, Netscape, AOLServer, with modular support for Servlet 2.2 and
any future version of the spec - you should get involved and be assured
that your code will be respected.
Even if it is not perfect, it lacks documentation, or have bugs -
I think we proved more than enough that your ideas matters, and the code
can be cleaned, documented, fixed and made part of the evolution.
Take a look at Tomcat3.0 - it is really bad code, great design ( no
documentation ). Take a look at Tomcat3.3 - it's the same design + ideas
from many other people, faster, cleaner, with some documentation. In the
process we had 2 releases that can be used in production sites.
Yes, it wasn't easy - it's easier to throw away everything and start with
your own ideas. But it's other's people ideas that matter, and it's worth
spending the time to understand them and reuse their code. And there is a
chance others will do the same with your code and ideas. That's my
understanding of open source.
Costin
