Related articles |
---|
[2 earlier articles] |
Re: Need advices for creating a proprietary language sandy.harris@sympatico.ca (1998-03-06) |
Re: Need advices for creating a proprietary language joachim.durchholz@munich.netsurf.de (Joachim Durchholz) (1998-03-07) |
Re: Need advices for creating a proprietary language mc@hack.org (MC) (1998-03-07) |
Re: Need advices for creating a proprietary language markh@usai.asiainfo.com (Mark Harrison) (1998-03-07) |
Re: Need advices for creating a proprietary language nixon@softlab.se (Leif Nixon) (1998-03-12) |
Re: Need advices for creating a proprietary language tucny@km1.fjfi.cvut.cz (Ondrej Tucny) (1998-03-12) |
Re: Need advices for creating a proprietary language ddd@hplisolx.grenoble.hp.com (Dupont de Dinechin Christophe) (1998-03-15) |
Re: Need advices for creating a proprietary language byte@lmn.pub.ro (Laurentiu Badea) (1998-03-18) |
Re: Need advices for creating a proprietary language kochenbu@khe.siemens.de (Andreas Kochenburger) (1998-03-30) |
From: | Dupont de Dinechin Christophe <ddd@hplisolx.grenoble.hp.com> |
Newsgroups: | comp.compilers |
Date: | 15 Mar 1998 00:25:24 -0500 |
Organization: | Hewlett Packard Cupertino Site |
References: | 98-03-039 98-03-107 |
Keywords: | design, practice |
> Lyn A Headley wrote
> >Writing a compiler is very hard. If you have never studied the
> > ....
Ondrej Tucny <tucny@km1.fjfi.cvut.cz> writes:
> Are you sure? I think writing a compiler is one of the easiest things,
> I've ever done. I think you need about one week to practise the basic
> algorithms and then you should be able to write a simple compiler. But
> yes, you're true if you're writing something like Ada. I wrote a
> compiler of a language may be nearsy as complex as Ada and it took me
> two years.
[Rest of the description of a simple compiler deleted]
I kind of agree with Ondrej, except I would add the following comment:
"The difference between theory and practice?
In theory, there is none.
In practice, there is one."
Now, writing a simple toy that parses expressions and generates an RPN
expression is really easy, that's true. As is the development of an
algorithm to add two one-digit positive integers. For a real compiler,
that's another story, as it is another story to implement an efficient
floating-point division in hardware :-)
However, the directions given by Ondrej really make sense as a good
starting point. But if you are really interested in the subject, there
are tons of books out there that you should consider reading (who said
"The Dragon Book" ???)
--
Return to the
comp.compilers page.
Search the
comp.compilers archives again.