Like a regex, but readable?

Question

Not to raise the ire of uncle_alice, but try as I might to read regex
tutorials, any sufficiently long regex pattern string is indistinguishable
from magic to me. It's unreadable.
So I'm wondering if code that step-by-step constructs a regex object,
without resorting to string symbols, would be more comprehensible?
Are there any example of this being done in any library? Or should I
just drink the Kool-Aid?

Answer 1

Actually, I agree: regexes cram too much meaning into too few characters, and end up looking like chicken scratch. Also, they were originally designed to match "regular" grammars (whatever they are), while we usually use them on non-regular grammars like natural languages, programming languages, HTML, etc.. Thanks to that impedence mismatch, some tasks that are conceptually simple are difficult to perform, if not impossible.
Trouble is, there's nothing else out there that's as powerful, general, and convenient as regexes. The only thing I know of that could completely replace regexes is a parser generator like Antlr or JavaCC, but those are way too heavy for most tasks. There are projects, like the two listed below, that try to make it easier to generate parsers, but they still lose out to regexes on the convenience front.
http://bnf-for-java.sourceforge.net/
http://jtopas.sourceforge.net/jtopas/index.html
I'm afraid the only thing that's going to be able to supplant regexes are better regexes, like the ones the Perl community is working on:
http://dev.perl.org/perl6/doc/design/syn/S05.html
But Ipthar help the schmuck who gets the job of shoehorning those into Java! ^_^