I dearly love revising old programs. My sense of style in programming has evolved over the years as I've garnered experience, first with C++, then with Perl, and now primarily with Python for many a year.
I despise revising old documents. Reading them triggers memories of my intention in writing them and the clarity of those memories makes it difficult to find better forms of exposition.
It seems to me as those two impulses should be more consistent. Part of the incompatibility is the gratification of programming (when it is going well) is external whereas revising old written documents usually offers only a solitary or deferred gratification.
Part of it is the sense of a clear path when improving programs; making them more modular and thus easier to follow. Renaming symbols to more clearly indicate their meaning. Improving the validation of program logic and more clearly indicating exceptions where the program is expected to fail, all of these things bring a tangible sense of improvement.
Clarifying, making more rigorous, extending previous written exposition: all of these things are judgment calls that once I have made them and time has passed I am loath to overturn. Oft times I will start from scratch and see if something better can be created.
Comments
Post a Comment