Path: csiph.com!v102.xanadu-bbs.net!xanadu-bbs.net!news.glorb.com!news-out.readnews.com!news-xxxfer.readnews.com!news.misty.com!news.iecc.com!nerds-end
From: compilers@is-not-my.name
Newsgroups: comp.compilers
Subject: Re: Good practical language and OS agnostic text?
Date: Thu, 19 Apr 2012 19:05:16 -0000
Organization: Compilers Central
Lines: 57
Sender: news@iecc.com
Approved: comp.compilers@iecc.com
Message-ID: <12-04-044@comp.compilers>
References: <12-04-027@comp.compilers>
NNTP-Posting-Host: news.iecc.com
X-Trace: leila.iecc.com 1334892497 87840 64.57.183.58 (20 Apr 2012 03:28:17 GMT)
X-Complaints-To: abuse@iecc.com
NNTP-Posting-Date: Fri, 20 Apr 2012 03:28:17 +0000 (UTC)
Keywords: books
Posted-Date: 19 Apr 2012 23:28:17 EDT
X-submission-address: compilers@iecc.com
X-moderator-address: compilers-request@iecc.com
X-FAQ-and-archives: http://compilers.iecc.com
Xref: csiph.com comp.compilers:587

gah@nospam.ugcs.caltech.edu wrote

> You might be right that all books are one (or more) of those three,
> but then you should choose from among those.

If that would have been an option I would have done it by now.

> C is a nice, simple language in which to describe compiler design, and
> also not so bad a language in which to write compilers.  Given that,
> it isn't a bad start toward writing one for another language, and/or
> written in another language, unless you don't know C.

I don't know C and it's not a good choice on z/OS. I am familiar with
assembler and I used to know PL/I fairly well and could come up to speed
with it and use it if it would be a good implementation choice. I don't
think it would though, because it requires the IBM runtime which is licensed.

> Both C and Java are fairly simple, and reasonably similar, with much
> of the complication moved to the library. That means a complete
> compiler can be described relatively simply, covering all the
> important ideas. Once you learn those, you will be ready to go on to
> other languages (both for compiling and writing the compiler in).

I really don't like C or Java and I didn't come out and say it because I
don't mean to start a holy war. I have looked at them. C just doesn't have
much value on z/OS. Java is too limiting in other ways. I am familiar with
the older languages and never was interested in much of anything that came
later although I have some experience with modern scripting languages.

> You don't say what language you are interested in.

I am not exactly sure. I was considering a subset PL/I or PL/M variant or
maybe even a new language. Even a super Pascal or Modula-something would be
interesting to me.

> There are some complications to writing Fortran compilers not covered in
> most books. Both Fortran and PL/I don't have reserved words, which
> requires special handling by the compiler. Fixed form Fortran ignores
> blanks, which requires a special lexical analysis technique, but most of
> compiler theory is in parsing, which is reasonably language independent.

I think Fortran would be hard because I couldn't write the libraries needed
due to my lack of mathematical background. F77 would be an interesting
project, the latest Fortran is way more complicated than I would attempt.

PL/I is interesting because I have access to old and new PL/I compilers and
good doc for z/OS. I figured I could develop the grammar from the manuals
but I know it would be a huge project, more than I have the time and ability
to do and a better PL/I than I or many people could write already exists,
so there's not much point in it.

> It seems to me natural that a pseudo-code description will be more
> theoretical, and harder to understand, than one using a well-known
> high-level language.

That's a bit of a tautology. If the well-known language is something /you/
know then yes! Otherwise..