5 ms·
Why Lisp macros are cool, a Perl perspective (2005)
- leephillips 8y agoPretty good read, but it's from a mailinglist, and is littered with quoted-printable residue. If someone wants to preserve this, it would be worth converting it to a readable form.
- forgotpwd16 8y agohttps://ptpb.pw/p3c5 https://ptpb.pw/p3c5
- leephillips 8y agoThanks!
- rurban 8y ago>Maybe you would like this to work in Perl: > sqrt($x) = 12; >and now $x contains 144. No, too bad. That is impossible, or at least extremely difficult. (Maybe you could do something ridiculous, like replacing the built-in sqrt() with a user-defined "lvalue" sub that tied its argument or something. Oy.) Adding variables to the lhs of an equation, logic programming, is something I'm toying with the last years. I wanted to name it cps, but now use logic sounds better to me. Every primitive op knows about its inverse function, and assignment is a matcher, as in functional programming. perl unfortunately has no matcher, their smartmatch op cannot match structures, it really is a dumbmatch. but that's easily fixable with a few lines of recursive code. there do exist nice prolog systems for lisp, and even a slow one for perl. Currently I'm using picat for such tasks, mixing declarative logic with normal forward expressions. http://www.picat-lang.org/ http://www.picat-lang.org/
- lizmat 8y agoIn Perl 6 this is almost trivial: sub sqrt($scalar is rw) { Proxy.new( FETCH => { $scalar }, STORE => -> $, \value { $scalar = value ** 2 } ) } my $x; say sqrt($x) = 12; # 144 say $x; # 144 I guess this could use some syntactic sugar, but the basic components to do this are readily available and simple to implement.
- rurban 8y agoPartially (that's the perl lvalue solution), but the problem shouldn't be solved by cheating, but by implementing a symbolic logical solver, like computional algebra sytems (Maxima, Sage, ...). To free $x on the left side square both sides. { use logic; sqrt($_) = 12; print solve($_); } => $_ = square(12) = 144; This is usually done by providing prolog-like rules for each and every operation, where many ops could provide more than 1 solution. Then feed them into a solver for more speed. Like with picat or cbmc. Lisp is perfect for that. With perl it gets hairy, but still should be possible.