Re: Languages: The Bigger the Uglier (Martin Neitzel)
23 Feb 1996 18:30:53 -0500

          From comp.compilers

Related articles
Possible to write compiler to Java VM? (I volunteer to summarize) (Peter Seibel) (1996-01-17)
Re: Aliasing in ISO C (1996-02-16)
Re: Languages: The Bigger the Uglier (Pat Terry) (1996-02-20)
Re: Languages: The Bigger the Uglier (Dr A. N. Walker) (1996-02-23)
Re: Languages: The Bigger the Uglier (1996-02-23)
Re: Languages: The Bigger the Uglier (1996-02-23)
Re: Modula II (Pat Terry) (1996-02-26)
| List of all articles for this month |

From: (Martin Neitzel)
Newsgroups: comp.compilers
Date: 23 Feb 1996 18:30:53 -0500
Organization: Gaertner Datensysteme, Braunschweig, Germany
References: 96-01-037 96-02-187 96-02-237
Keywords: modula, standards, design

The Modula-2 definitions are indeed relevant to the points made.

The original Wirth M2 report (in "Progr. in Modula-2") contains
very irritating uses of footnote-sized fine print. For example:

10.1. Formal parameters
.ps -2
Restriction: If a formal parameter specifies a procedure
type, then the corresponding actual parameter must be either a
procedure declared at level 0 or a variable (or parameter) of
that procedure type. It cannot be a standard procedure.
.ps +2

It is annoying enough to dig the fine print in insurence contracts.
To discover the same method applied to language definitions gave me
the creeps. (I had to write an M2 interpreter.)

The whole PIM2 report is written in a superficially formal and
legalese style but has in fact an awful lot of unclear/undefined
things. (The often criticized K&R reference for C on the other hand
is very clear on the intentions of the authors and does a much better
job to give clear rules.)

While I could quote PIM2 by heart, I just had occasional contact with
the M2 ISO draft in VDM. But I can definitely say that it answered my
implementor's questions very clearly and very trust-worthy. At least
I felt very confident to have done complete investigations of the
questions I had. The cross-references and formal definitions make it
very easy to find all places of relevance - and to be sure about it.
I could cover a topic _reliably_ much quicker in the 700 pages ISO
draft than in the 30 pages PIM2 report.

> [I don't suppose that anyone in the Modula-2 committee took the failure of
> the VDM effort as an indication that the language was too big? -John]

As far as I'm concerned, they did a great job. Modula-2 has really a
mess of rules hidden in those 30 pages report. It's no wonder that
that finally shows in a _long_ formal spec.

If the Committee got tired because of self-induced featuritis, then
that's their problem. As an implementor, I'm happy to be able to read
the succintly defined outcome and to stay awake. Wirth did such a
lousy wishy-washy job that it made me lose my mind at times.

Martin Neitzel

Post a followup to this message

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