>>>>> "DS" == Dan Sugalski <[EMAIL PROTECTED]> writes:
DS> Actually, I'm expecting almost nothing to be written in pasm after
DS> we get a working parser and compiler. We might write template code
DS> for the compiler in it, but that's probably about it. Well, other
DS> than the JAPHs that aren't being written in APL or Intercal or
DS> something. (I can see someone writing a befunge parser just for
DS> the wacky JAPHs possible...)
i envision some code being written in pasm for the same reason any code
is written in assembler, speed and fine grained control. there may be
reasons not to use perl6's version of Inline::C and pasm is a good
compromise. some inner loops or short subs or special glue code (thread
semaphores, language specific API's, etc.) could be hand coded by some
lunatic genius (and we have a few of those in our midst). not that i
recommend doing this but if we get a true macro assembler with decent
power and ease of use (and by writing it in perl5/6 it will be easier to
implement), there will probably be some hand coding of pasm. we have
also speculated about some perl6 functions being written in pasm as well
or maybe code generated and then hand optimized. so let us assume there
will be pasm code being hand written and support it now. we can always
add a macro assembler as just a new front end whenever we want.
uri
--
Uri Guttman --------- [EMAIL PROTECTED] ---------- http://www.sysarch.com
SYStems ARCHitecture and Stem Development ------ http://www.stemsystems.com
Search or Offer Perl Jobs -------------------------- http://jobs.perl.org
