Weekly Report 22nd through 28th Feb

This week we were able to start on the project design.

Meetings

The team met on Monday, 22nd Feb. We had a session in which we discussed topics we were supposed to be learning. Shyam talked about RCS and we decided to use it in controlling our source code, when we reach that point. He also sent out a mail describing commonly used commands. John talked about fltk and fluid and we discussed ui layout and how the user interacts with the interface. We seem to have all the parts necessary to build the interface now. Sada described the back end of our system, the scheduler. We compared different algorithms to do this, and currently decided to go with the simplest of these.

Design

Broadly the system has been divided into two halves. The interface part that the user interacts with and the scheduler, the back end that actually executes the data flow. These two halves communicate through a file which represents the dataflow to be executed. We will require a parser to read this file format into a common data structure and routines to generate it. The modules that the user interface uses are described to it in a "library" file. we will need another routine to read in this file format. As the user connects/disconnects modules in the interface, the interface data structure that represents the dataflow is modified. When the user selects to run the dataflow, the interface writes out the file which is read by the scheduler and executed. Havent started work on the implementation manual as yet.

We still seem to be on schedule. Need to get started on the implementation manual though.