>One approach to development in CL is to evaluate a form you would like to have. You end up in the debugger. You write the implementation until the error goes away.
This doesn't sound very sustainable or effective to me... almost all the bugs I encounter in programming are algorithmic in nature. The code compiles without error and runs without crashing, but it just happens to do X instead of Y. The code fails because I didn't understand the problem well enough. That's one thing that computers can't really help with yet.
I think part of the idea is that you write the implementation, test it immediately to make sure it does Y, then restart.
Note that this is not how I really program either, but I've not tried it enough to have an opinion about whether it's effective forme, and I suspect it'll be one of those 'matter of taste' things (I barely use interactive debuggers at all).
This doesn't sound very sustainable or effective to me... almost all the bugs I encounter in programming are algorithmic in nature. The code compiles without error and runs without crashing, but it just happens to do X instead of Y. The code fails because I didn't understand the problem well enough. That's one thing that computers can't really help with yet.