It seems that Lars has not understood my comment correctly and mispelled my name, too :(
Open Source and software design fit very nicely together. In fact if you look at many good open source projects, like Cocoon, Tomcat, Xerces, Xalan, Avalon to name only a few you will find a very good software design. The problem is how to reach this design?
The developers are on different locations. If one developer is currently sleeping, two others have day-time and are active. There is often no other communication channel than emails. But for example the Avalon team has reached a very good design simply by exchanging emails. So instead of exchaning UML models or something like that, text descriptions and java interfaces were directly discussed. This works very well.
My statement is, that if there is currently no real need and noone willing to implement the design later on, it will result in endless discussions. This is often the case for RTs (= Random Thought) and RTs was the topic I was talking about. But if you start a new project from the scratch, you can discuss it by email and implement it. Then it works because there is a real need to have this software.
And first doing something and discussing this later on is a good software technique as well: it's called prototyping. I personally believe that the best software design can often only be achieved by doing a prototype and then seeing what's wrong there.
1:25:46 PM 
|
