Mark Lowe wrote:
What are the exact reasons you don't like a source
distribution?


If I were working on my own then there's not a huge problem, but this
isn't the case. There are several agencies involved and the usual
political difficulties when changing things like build files and
versions.  Someone walking into this to do a request fix. but has no
previous knowledge of the application gets bogged down, having to
understand that cocoon needs building, the build scripts have grown in
complication to the point of being unmaintainable.

Being able to have a cocoon based project, define the dependencies in
the project.xml (or even an ant build file) build your source code and
webapp resources into a war, would be a huge step in the right
direction. Its the little things that count.Someone on the original
thread mentioned the classic ./configure;make with
./build.sh;:/cocoon.sh , isn't that the point?

Mark,

cocoon is a framework that in real life feels like a huge servlet. It's a pain. It bugs us, it bugs you, it bugs everybody.

We had to make decisions, to solve the needs, decisions that were maybe wrong or maybe they were half-baked, or hoped in events that didn't happen.

Cocoon has the highest number of dependencies of any other open source java project in the world. This is a fact. An honor, but also a burden. Our build system is a complete nightmare... 150kb of ant build... Several of the Ant features came out from *our* need for coocon and I personally implemented them.

We have been working on designing the block system that would allow you to use cocoon just like a black-box: deploy and forget.

Such a system is already designed, we call it 'real blocks' but it's very complex and very ambitious, since it solves all the problems on the table at once.

But it had to happen all at once (no development incrementality in its design, we didn't think of that when we started!) and nobody was able/had enough energy to do it.

The death of the Avalon project unleashed our will to maintain our own container, this unlocked a lot of potential that had been locked for years.

The community is working actively to solve these problems. Several steps have been made and very recently Reinhard is pushing a lot on making it happen... also because we now understand that it has to happen incrementally, so we are willing to sacrifice elegance for anything that moves us forward.

This is an open project and everybody is welcome to participate in the way they find more appropriate.

But you have to understand that the community reaction, as well, will be driven by the above attitude.

You can also decide that this project is not a good fit for you. We won't try to convince you of the opposite.

--
Stefano.



Reply via email to