Hello I think we are mixing things up. Well designed and documented code is great and the easiest thing to achieve that is writing small applications. But MMBase consists of many smaller parts that often where only partially implemented(just to keep things simple?). Do you have any idea how much work has do be performed because somebody once choose to only implement what he needed without thinking of the future. How much work has gone into refactoring code while keeping it backwards compatible. Do you have any idea why we do not have transactions? good import /export features? working editors? and finally documentation?
I think that it's because every time you start writing code you end up in a totally different part of mmbase asking yourself how that code could ever has worked. You have to know the MMBase bible to avoid the pitfalls. There is currently simply not much room for inovation > An Extreme programming rule: Do The Simplest Thing That > Could Possibly Work. I am a strong believer of the "do one small thing good" that's not the same as " Do The Simplest Thing That Could Possibly Work" > These future projects will show where mmbase is not extensible and > what has to be changed to fit the need. Don't we already know enough? We are lacking a good vision , but don't expect some external party to create something like the function framework. We are part of the community and I don't think or goals is to provide something that "works4me". The real problem of course is that in MMBase nothing is small. To make the editwizards work for example you need so much and it's a really miracle that they still work. (not really since there is so much effort going into them) > Something being logical is based on earlier or known statements or > conditions. Hard to read source code with too many features will redefine > being something logical to being wrong with confidence. > The things that are logical now are not always logical in the future. The > ways in how mmbase is used will change over time and new ideas will show > that things aren't as logical as we thought they would be. How these changes > are managed will make mmbase successful or not. Something that is logical is easy to understand and easy to explain. Even if the usage changes over time you still can understand it. Whatever we build should be ready to be used by some other geek. > I replied on Michiel's suggested "change" to leave it off the todo-list, > because it has no return on investment at the moment. I used a very big word > (over-designed) to emphasize this. From the time I started using mmbase up > to now the handle field seems okay for me and according to the code it is > reasonable too. So you have been happy all the time :) I have never been happy about the handle field. The first thing I tried when i started using MMBase was to create a builder with 2 images. and that was not possible at the time :). _______________________________________________ Developers mailing list [email protected] http://lists.mmbase.org/mailman/listinfo/developers
