Kees Jongenburger wrote:
I might not be the person who everybody should listen to
but I have the feeling that everybody has been working very hard
(storage,editwizard(how hard),taglib,fuctions) but that we could do much better
if there was a bit more thinking and evaluating in the process. Even
a project manager would have seen that the editwizards was a waste of
time.
I do agree that there is a lot of hard work being done. The editwizards however are project horribly gone wrong because of 'featurism' the parties involved wanted way too much, and also in very differing directions.
But if michiel want's to improve the code bit by bit I have no problem what so ever, i am even very very happy (but don't ask me what I think of it if you only want positive feedback). We don't need more directions , we need better directions. I am not happy at all about the way code is moved to applications, it looks almost random and I have posted enough on the subject (just get enlightened and read the ideas about maven). And that is even worse than Michiel's hacks because the cleaning project is a project, we so just can' t stop it.( master gomez is the master )
Well I don't consider the moving out of some functionality 'random'. I think one of the goals is to make the core more manageble, with clearer dependecies.
I haven't heard much from you lately. Don't you want to say for example that in mmbase 0.9 the builder configuration was in the packages and that it was moved to a config directory[because...], and now some people think again that dtd's and configuration belong in the the packages? Don't you care about the direction als long as the steps are small?
I don't mind centralizing the point were all config files are kept. I do object if it gets to a place where it is difficult to change and is dictated by the appserver.
What's you feeling we should be doing? How would you like to See the development process going?
I think we need better directions as you said. Part of that problem is something about which the MMC discussed yesterday, namely a sort of roadmap/featurelist in which we give a more general pointer where we are going. Secondly I think we should minimize 'big' projects as we are seeing that 'big' projects take forever to complete as things are changing under their footing. So I am more for 1 big project every release cycle, and a bunch of small ones.
And I would like it if some things are 'set in stone' so that they don't change to much. Examples are the database layers and the editors, there have
been problems with all versions we have and have had. However I think it
would be best if we don't always strive for the 'best' solution but
occasionely for the 'second best' if it gives us stability and progress.
As I said before I have a strong feeling that some people are more interested in making MMBase 'beautiful' codewise then focusing on actual improvements. Also I think some people are thinking of MMBase as more a playground instead of a software package in production use.
If we follow a road with more focus on that, we probably get a more stable core and can focus on more exciting enhancements to MMBase technology wise.
-- Rico Jansen ([EMAIL PROTECTED]) "You call it untidy, I call it LRU ordered" -- Daniel Barlow
