Re: Language Books

mslamm@olive.mscc.huji.ac.il (Zvi Lamm)
24 Jul 1998 12:25:57 -0400

          From comp.compilers

Related articles
Language Books rrader@ford.com (Russell Rader) (1998-07-20)
Re: Language Books Juergen.Kahrs@t-online.de (1998-07-24)
Re: Language Books mslamm@olive.mscc.huji.ac.il (1998-07-24)
Re: Language Books bwb@concentra.com (Benson, Brent) (1998-07-24)
Re: Language Books mslamm@olive.mscc.huji.ac.il (1998-07-24)
Re: Language Books tenger@idirect.com (Terrence Enger) (1998-07-26)
Re: Language Books carsten.link@cityweb.de (Carsten Link) (1998-07-26)
Re: Language Books cbrtjr@ix.netcom.com (Charles.E.Bortle Jr.) (1998-07-26)
Re: Language Books rlrader@ix.netcom.com (Russell Rader) (1998-07-26)
[3 later articles]
| List of all articles for this month |

From: mslamm@olive.mscc.huji.ac.il (Zvi Lamm)
Newsgroups: comp.compilers
Date: 24 Jul 1998 12:25:57 -0400
Organization: Hebrew University, Jeruslem, Israel
References: 98-07-131
Keywords: books

Russell Rader (rrader@ford.com) wrote:
:>Can anyone out there give me your recommendation on books on programming
:>language design? (not compiler design itself, but overlap between the
:>two would be ok, I guess) I went out to amazon.com and clbooks.com, and
:>there were actually too many choices. But I don't want to spend $50 on
:>a book based on a one paragraph blurb.


:>[Language design has always been appropriate for comp.compilers so long as
:>it has some relation to compiler design. No arguments about where the
:>semicolon goes, though. -John]


This is a lovely topic, but one which I am afraid isn't dealt with
extensively. You may find Jon Bently's work on "little languages"
apropriate, and some of his other works.


One thing that I'd like to say, before in depth compilation issues are
raised, is that one must contrast between the compiler's viewpoint (or
let's say the compiler's writer viewpoint) and the viewpoint of the
average programmer. We tend to confuse the two, saying that languages
easier to parse, are easier to understand. This is sometimes the case
(for example fixed precedence order vs. user defined operator
precedence), but this isn't some "universal law".


You should, perhaps, browse Chomsky (e.g, Aspects of the Thoery of
Syntax) for the human standpoint. One of the major contributions
Chomsky made was to use parsing theory to dismiss certain classes of
grammer as candidates for humam language grammers. Like saying this
lang. contructs is allowed in english, but can not be in CFL. This can
be used "backwards" when thinking about human languages.


You should esp. enjoy Chomsky's distinction about the "practice" of
grammar use as opposed to the theoretical power of the grammar class
itself. If I remember correctly (the book isn't here) this is in
chapter 3.


As a last note, please consider:
1. What is easy to parse
2. What is easy for a human to parse
3. What is easy for a human to understand
4. What is the semantic structure of the language (which has nothing
tying to grammar)


Note that 'strange' languages (e.g, APL, Prolog) usually have strange
syntax. However Icon has a strange execution model but has a pretty sane
syntax.


Sorry for the long winded reply, but this is a lovely topic...
--
Ehud Lamm mslamm@pluto.mscc.huji.ac.il
--


Post a followup to this message

Return to the comp.compilers page.
Search the comp.compilers archives again.