XPost: alt.folklore.computers
From: Peter@Iron-Spring.com
On 1/5/26 17:57, c186282 wrote:
> On 1/5/26 13:49, John Ames wrote:
>> On Sat, 3 Jan 2026 08:31:33 +0000
>> The Natural Philosopher wrote:
>>
>>> "The statement "Pascal has no I/O" originates from
>>> Brian Kernighan’s 1981 essay, "Why Pascal is Not My Favorite
>>> Programming Language".
>>>
>>> Kernighan argued that the original 1970 definition of Pascal was
>>> severely limited for systems programming because:
>>
>> Yeah, that was it - not *no* I/O in the sense that was true of Algol,
>> but weird and constrained in ways that betray its origins as a teaching
>> language. Mainly, files are assumed to be of a uniform structure; you
>> can have a FILE OF CHAR or a FILE OF INTEGER, but not a file containing
>> both strings and integers. If you want to do *that,* you're supposed to
>> make a struct and have a FILE OF that, but this too has to be the same
>> across the whole thing. Files of mixed or variable structure? Who uses
>> *those!?*
>>
>> Like many of Wirth's design choices, it sounds simple on paper but is
>> unnecessarily confining in the Real World - and, as Kernighan points
>> out, there were no "escape hatches" for extending the language from
>> within, leading to a bunch of proprietary and mutually-incompatible
>> variants. Obviously, it's been decades and the landscape has changed
>> substantially, but it really was dunderheaded at the time.
>
> Wirth was an 'academic' - and Pascal/M2/M3 kind
> of reflect that.
>
> However it WAS easy to extend the language - add in
> those Real World necessities. By the time Turbo Pascal
> hit the scene there really wasn't anything you could
> not do with Pascal.
>
> And I still write in Pascal fairly often - like
> it better than 'C'.
>
I'm not sure to what extent there was an attempt early on to standardize
the extensions, but this would have helped adoption of the language
immensely.
--- SoupGate-Win32 v1.05
* Origin: you cannot sedate... all the things you hate (1:229/2)
|
