The main problem in my opinion is that most of the Open Source Software is developed in a non-structural manner. This has as a result that the developers come up with something totally cool, but they forget about the simple useful stuff.
An example: For a little project we had to work with a database. That was all fine, but the db was open source. Nothing wrong with that in itself either. But now comes the problem, the db had several very cool thing (like the ability to create a new procedural language within the program), but simple thing were missing and overlooked. We had to struggle our way around very nasty problems (Triggers which could only call opaque functions, which on their turn couldn't be created).
The thing we're looking for in the Open Source community, is a good engineering model to be used. As a matter of fact, developing open source software should happen the same way as closed commercial software is developed, without the time constraint. SO beginning with requirement, specifications etc. would helkp a lot in my opinion. I've seen this for my self. A of course: Go and talk with John Doe about what He really wants to see, and implement that before you create the very cool features.
To explain this post a bit. Thanks to the great guys of SNT, CIV and the ones named above, the uplink is back here. This is realized by setting setting up 4 sattelite dishes, and now there is a uplink through these dishes. It still isn't as fast as it used to be, but at least we're back online and much faster than was foreseen this morning.
Slashdot Mirror
The main problem in my opinion is that most of the Open Source Software is developed in a non-structural manner. This has as a result that the developers come up with something totally cool, but they forget about the simple useful stuff.
An example: For a little project we had to work with a database. That was all fine, but the db was open source. Nothing wrong with that in itself either. But now comes the problem, the db had several very cool thing (like the ability to create a new procedural language within the program), but simple thing were missing and overlooked. We had to struggle our way around very nasty problems (Triggers which could only call opaque functions, which on their turn couldn't be created).
The thing we're looking for in the Open Source community, is a good engineering model to be used. As a matter of fact, developing open source software should happen the same way as closed commercial software is developed, without the time constraint. SO beginning with requirement, specifications etc. would helkp a lot in my opinion. I've seen this for my self. A of course: Go and talk with John Doe about what He really wants to see, and implement that before you create the very cool features.
To explain this post a bit. Thanks to the great guys of SNT, CIV and the ones named above, the uplink is back here. This is realized by setting setting up 4 sattelite dishes, and now there is a uplink through these dishes. It still isn't as fast as it used to be, but at least we're back online and much faster than was foreseen this morning.