![[Home]](http://www.bayleshanks.com/cartoonbayle.png)
clojure's :as
e.g.
(defn make-rectangle [[p1 p2 p3 p4 :as vertices]] (let [a (distance p1 p2) b (distance p2 p3)] (assert (= a (distance p3 p4))) (assert (= b (distance p4 p1))) { :kind :rectangle , :vertices vertices, :a a, :b b}))
i think it is for binding part of a pattern while also binding the parts of that part
...
clojure's :keys
--- unlike regexps, a pattern matching language must match over graphs that can have loops
however, i bet you could render a regexp-like pattern matching language always terminating (total) by simply detecting when you are re-traversing a loop while trying to match the same sort of thing (this would allow you to terminate that sub-match with either success or failure depending on whether the match condition is forall (success) or exists (failure))
---
allow bounded FOL, e.g.
and, or, not, and any/all over finite structures
or perhaps we should allow any/all over infinite structures but only when the compiler can statically infer whether the any/all is true
--
http://blog.sigfpe.com/2010/08/constraining-types-with-regular.html
--
perl6 may have nailed string grammars
i think its implemented here, in intermediate language NQP:
https://github.com/perl6/nqp/tree/master/src/QRegex
---
here are OMeta like things for Ruby:
" The closest thing to Perl6 grammars in Ruby would be the Ruby-OMeta project (make sure to also take a look at Ryan Davis's fork), which unfortunately ist still under development. (Or rather, no longer under active development.) "
https://github.com/Aquasync/Ruby-OMeta/
https://github.com/ZenSpider/Ruby-OMeta/
---
http://www.tinlizzie.org/ometa-js/#Sample_Project
http://www.tinlizzie.org/ometa-js/#OMeta_Tutorial
https://www.google.com/search?client=ubuntu&channel=fs&q=PyMeta&ie=utf-8&oe=utf-8
---
in OMeta, matching list data is like PCRE regexps which always end in $, that is, the whole thing must be matched. This seems easier to think about, but is perhaps inefficient in cases where you are just ending the match with .*? but i guess the compiler could in theory figure that out.
--
