30/03/08 - Framework and tools.

Admittedly, I posted the previous post five minutes ago, though it should have been posted on Friday. I got home late on Friday, so never got a chance and yesterday... yeah. Anyway, on to todays real post:

I didn't really do much since Friday, mainly because of time constraints - unfortunately I have a lot lot more needing doing than just the project.. I did, however, get to plan out the featureset for the application editor some more and have also decided upon an architecture for the integration of the application editor and the message router/state machine.
I have decided to merge both programs. If I had more time for the project, I would probably keep them seperate, as that allows for greater flexability in the long run, but since the deadline is fast approaching, I decided it was more important to simply get it running. By merging them, it also makes it a lot easier to expose functionality between the two - because they are now one and the same! The application editor also benefits from already having the Twisted networking code implemented in the message router and it now has direct access to the state machine. In the morning, I hope to merge the two and refactor the resulting code to ensure there is no decrease in code quality. I expect this program to be an integral part of the framework, not only being the central point for users to access the frameworks built-in functionality through a grpahical editor, routing messages between the various daemons or haandling states and state transitions, but also providing a platform for python code to access the frameworks features for custom coded applications, which are not possible through the editor alone - and these applications should be able to add tools to the editor, for convenient access!
Hopefully I can manage to complete the code merger in a single day, in which case there should be another post here tomorrow. This will leave me to implement features on tuesday and wednesday. The goal now is that any core functionality is complete for Thursday, so that we can then build a few simple demonstration applications. I hope the demo apps will prove the framework and application editor to be intuitive and easy to work with. These demo apps can then be also used for our guinea pig testing :-)

If the rest of the project was a success so far, then the demo apps should be trivial to implement. After all, that is the whole point of implementing a framework.


A quick look at the calendar and schedule shows that we are still on track, though care must now be taken to ensure we don't fall behind. We need to have the demonstration applications complete within about a week and a half, meaning the framework needs to be in a useful state before then. That leaves us with a week or two for testing and evaluation and a week for documentation. The schedule should really be revised, since I don't think we are able to do antyhing after April. As it stands though, I believe we are about where we should be, perhaps a week behind, but nothing we cannot recover from.
At the start, most of the work was done by Graham, since there was little that could be done without the hardware and code that interfaced with said hardware. Now that it's mostly complete, it's my turn to have more work to do. So, now Graham is mostly testing and debugging, while I'm writing a bucketload of code. Almost IBM standards - millions of KLOCS! Ok, maybe not, but I wrote more code last week than I have any other week since the start of the project!