Re: Pascal vs. linkers, was The History of the ALGOL Effort

Jürgen Kahrs wrote:
><snip>
>
> It was the lack of a standard library that really killed Wirth's
> languages. And Wirth's refusal to take part in the ADA development
> took away his last chance to influence the development. It was around
> 1990 that I felt C had won. <snip>


There were so many reasons for the failures of Wirth's languages that
it is difficult to understand how they became popular for significant
lenghts of time


1. Lack of separate compilation


2. Lack of standard libraries, particularly I/O.


3. An over emphasis on language simplicity. Yes a small language can
have the basis for 99%+ of the features a programmer typically needs.
But once you start dealing with code more than a few hundred lines
long, or with unusual application, you too often miss the features
larger languages provide.


4. A superficial approach to language definitions, perhaps driven by
his emphasis on language simplicity. His language descriptions
achieved much of their brevity by consistently omitting details
necessary for a truly portable language definition. Compilers from
different vendors too often differed consistently in their treatment
of corner cases, making portability between compilers for a language
difficult. One of the (many) reasons the Modula 2 standard is so
large.


5. A lack of continuity. Code could not be easilly ported between any
of Algol W, Pascal, Modula, Modula 2, Modula 3, Object Pascal, or
Oberon. Perhaps Wirth should have paid more attention to Algol 68's
emphasis on orthogonality.


6. A consistent lack of interest in any form of standardization. As
near as I can tell he made no attempt to affect the initial Pascal
standard (too restricted to be useful), the extended Pascal standard
(too different from any existing implementation), the Modula 2
standard (too long in creation and too large), or gave any support to
an effort to standardize Oberon. In contrast Ritchie and Stroustrup
provided significant inputs to the C and C++ standardization efforts.

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