I spent the day working on pydap today: fixed a bug with the Oracle connector in dap.plugins.sql, added MSSQL support (though untested); and moved the 2.3 branch closer towards a release: I ported the CSV and Matlab handlers from 2.2 to 2.3, added some Javascript to the HTML download form.
I really hope I can release 2.3 soon… while the functionality is pretty much the same, I rewrote the code having in mind all the little things people asked me during the last couple of years. With 2.3, it should be easier to change the server without knowing Python, and to add functionality without touching the core. This way, I hope to make it easier both for users and for me to improve on pydap.
One thing that amazes me with pydap is that the number of lines of code keeps getting smaller and smaller since I started the project in 2003. The current base has less than 2.5k LOC (sic), for a complete Opendap client/server with support for NetCDF, Matlab and CSV files. Of course the code is small because I reuse a lot of 3rd party modules — some of them even written by myself — but it says a lot about the code that I have to maintain.
What’s left? I still need to write the documentation and make the new website. Add some more comments and docstrings. Check some old bugs. You know, the 20% of a software project that ends up taking 80% of the time. At least I learned to use unit tests on my code.
