From: Nikolai Weibull <mailing-lists.ruby-talk@...>
Date: 2004-11-19T04:10:48+09:00
Subject: Re: About Regular Expressions

* trans.  (T. Onoma) <transami@runbox.com> [Nov 18, 2004 16:50]:
> > > This is terrible.

> > Thanks.

> Nikolai the Terrible ...

Weird, my middle name is Iwan.  Except for the spelling, it all seems to
fit ;-).

> > Listen, my intent wasn't to place blame nor accuse any individual of
> > any wrongdoings.  I just wanted to point out that we'd been having a
> > lot of the same kind of discussions and that they seemed to suggest
> > a very poor understanding of regular expressions and their use among
> > the Ruby-Newby populace.
> >
> > Apparently I managed to press quite a few buttons with my post and
> > if anyone was offended I apologize.  I, however, still think my
> > thesis and conclusion are valid.
>
> Grace be upon you. It seems so few people are willing to apologize for
> anything these days. Thank you. I think the problem was that you
> weren't very specific about which posts you were referring and then on
> top of it came across a bit harsh.

Yeah, I realize that now...I always manage to do that for some reason.
It's not very nice to sound angry.

> > My reasoning was that people either weren't looking at the
> > documentation or that the documentation was too limited in nature.
> > Robert Klemme suggested that there was a lot of documentation
> > available (I agree).  If this is true, then why are people still
> > asking very basic questions?  Either they are too lazy to look up
> > the documentation, or they can't find any.  Both are a result of
> > there not being a good resource of introductory (nor advanced for
> > that matter) material about using regular expressions, especially
> > for Ruby.

> > I suggested that this situation be remedied, but that I couldn't do
> > it myself.  The general attitude so far seems to be that I should
> > shut the fuck up and write it myself if I think this is such a big
> > deal...

> I don;t think so. You may well have a point. So I welcome the
> outspoken concerns.

> My opinion is this: Regexp's are for CS-heads --those who really love
> programming for its it's own sake. Regexp are about as terse and
> cryptic as one can get, and thus have a steep learning curve.

Yeah, that's perhaps the biggest problem.  It's fine when you're working
with them interactively (well, I don't think so personally, but that is
one reason as to why they are terse to begin with (+ mathematics)), but
when writing them in a program, I think you should be presented with a
somewhat more understandable syntax.  That's what I'm working on with my
masters' thesis actually (but from the interactive side actually).

> Indeed, I have read a good bit of docs on them and I still have
> troubles. Rather than write more documentation, I think it would be
> the bomb if someone would invent the "Next Big Thing"(TM) in string
> pattern matching. But of course, that's a much taller order.

Yes.  It would be a welcome change though.  Perhaps proving P = NP
and/or English/Swedish/Japanese/Whatever = Context-Free would make
things more interesting ;-)

> P.S. Perl 6 regular expressions look to be a move in the right
> direction though. Has anyone checked that out recently?

Yes, it is a welcome change.  It promotes abstraction of regular
expressions as rules and sets of rules, called grammars.  Most of the
metacharacters are retained, but the end result is a much cleaner
language to work with (and a much cleaner implementation as well it
seems).  I still think they've gone a bit overboard, but they have good
reason.

Enjoy,
	nikolai

--
::: name: Nikolai Weibull    :: aliases: pcp / lone-star / aka :::
::: born: Chicago, IL USA    :: loc atm: Gothenburg, Sweden    :::
::: page: www.pcppopper.org  :: fun atm: gf,lps,ruby,lisp,war3 :::
main(){printf(&linux["\021%six\012\0"],(linux)["have"]+"fun"-97);}