Sunday, January 31, 2010

One other metric of progress

One way of measuring the progress of the compiler is how often I have to stop and debug it when writing code. My work pattern used to be - write a handful of lines of Tart code, figure out why the compiler was asserting, write a few more lines, repeat, etc. But today the interval between debugging sessions was more on the order of several pages of code rather than several lines.

Once I get to the point where I can write major parts of an application without having to stop and fix the compiler, it will be time to encourage other people to try it.

Another measure of progress is a general feeling of trust and confidence about the compiler. What I mean is, when coding, normally one doesn't spend any mental effort thinking "I wonder if this next line I write is going to reveal a bug in the compiler". Normally ones thoughts don't dwell at that level, and one can instead fully concentrate on the actual code being written. I'm afraid I am not quite at that level of confidence yet.

No comments:

Post a Comment