De : Maurice van der Pot <griffon26 kfk4ever com>
Envoyé : 30/10/2009 18:11:50 +0100
Objet : Planner-list predecessors for fixed date tasks
I totally agree ; remember my sentense just before your extract :On Fri, Oct 30, 2009 at 01:14:35AM +0100, Ricardo Fernandez wrote:is there any way to centralize a collective analyze and specification for algorithm evolution ? is there any conceptual documentation (exept code and its comments) for link rules and sheduling algorithm ? I have no time (like everyone) but should be interessed (probably like some others) to contribute to a conceptual specificationThe problem is not in coming up with an idea for scheduling, it is in finding someone who is willing to work on it and stick with it. When thinking about how scheduling should work it quickly pulls in many many other things that ideal scheduling functionality requires, such as split tasks. " the less restrictive is a tool, the more useful it is... and the more complex is the algorithm." and spliting tasks (for example) was to my mind an absolute necessity for schedulling taking ressources in account. ok for 'big ideas'.The bigger the ideas and requirements get, the more motivation you need to actually take on the problem of implementation. concerning requirements and precise specifications, instead, I thought it could help implementation as well as motivation. that's done !Additionally it is harder to motivate someone to implement other people's requirements than it is to get them to implement their own. In my opinion discussing requirements like this will go nowhere unless the one starting the discussion is the one who plans to implement it. Look through bugzilla for this feature, you'll see what I mean. and that's precisely why I suggested another way of work. in conclusion, I fear that those who hope Planner can become the accurate tool they need, there are two options : - discussion list ; a waste of time - waiting for developpers to find time and motivation... Regards Ricardo FERNANDEZ Regards, Maurice. |