In article <·················@pygmalion.mitre.org>,
····@pygmalion.mitre.org (Eric Peterson) wrote:
> P.S. Sonya K. warns us, but I think that it is worth repeating, that
> when specializing INITIALIZE-INSTANCE, an :AFTER method is probably
> the way to go. With anything else, trouble awaits the novice and the
> forgetful initiate.
There is [at least] one caveat to this: in some windowing code in MCL, it
is better to not use :AFTER, but to immediately use call-next-method.
Otherwise, one tends to get windows on the screen that are then visibly
filled in when the :AFTER method goes off.
> P.P.S. Another beginner's blunder might be considering CLtL2 the way to
> learn CLOS (I committed this one). I think that it's too detailed,
> theoretical, and implementor's-spec-oriented for a beginner. Sonya
> Keene's book, I think, is very good (I haven't read any others).
> After that, CLtL2 would be very instrucive.
Actually, for the experienced-hacker-learning-CLOS, I found CLtL to be
better, but maybe that was just my experience. I would certainly not
recommend going for any length of time in LISP without the bible, er, CLtL.
:-)
Particularly useful for the beginner is to read a section of Keene/
Winston/ whatever, and then immediately look up the relevant CLtL section
for all the details left out.
--
--
--
-- "TANSTAAFL" Rich ·····@ils.nwu.edu