No, I’m looking at Pascal from a 1978 perspective. If there had been a good C compiler for the PDP10, TeX would not have been written in, actually, a macro language on top of Pascal (to help work around its shortcomings both in string handling and memory management; oh and that statement labels have to be numeric).
From a 1978 perspective, Pascal was indeed already becoming dated, sure. But by then Wirth had already developed Modula, and started Modula 2.
But even from a 1978 perspective, most of your criticism of Pascal is for being bad at things it was never designed to be good at, and never pretended to be good at. It was a teaching language trimmed down (in terms of concepts) from Algol W for ease of implementation over feature-completeness, not designed to be a systems language.
That it became the basis for so many language implementations that opted to solve the limitations of Pascal over starting from scratch is a testament to its success.
But at the time, it was claimed that it was a systems language! No less than C. A. R. Hoare claimed in 1977 that it was "the best language in the public domain for purposes of system programming and software implementation."
Having had to use a "Pascal as defined by Wirth", by the way, everything in that criticism is dead on. Almost every one was an actual pain point for me when trying to write actual programs. (Not xor, but just about everything else.)
Hoare may have claimed that, but Wirth did not to my knowledge, and judging Pascal by the standards of praise given by someone who didn't design it is a bit weird.
(It's also worth considering that Hoare did posit e.g variants of Pascal, and it's unclear if he was writing of unextended Pascal, or in the abstract.or about a variant)
I don't think the criticism is wrong with the caveats Kernighan gave, because he went to lengths to address its use outside of the educational domain it was designed for.
Just not very relevant unless you happen to be dealing with the truly rare persons that were promoting an unextended Pascal as a systems language by 1981.
I am that rare person. I was using it as a systems language in 1986. Without the Turbo extensions. (We did have separate compilation.) Why yes, I am still bitter, does it show?
If you had separate compilation you were not using unextended Pascal. And it does not sound like you thought it was suitable, so you don't seem to fit the description?
Who in the world pushed a limited Pascal on you that late?
It does seem like something Wirth himself would have seen as an awful decision given that he'd worked on replacements for Pascal for a decade at that point, and wrapped up his work on Modula-2 to move on to Oberon.
I don't know who made the decision. It certainly wasn't me. But I only had a year of experience, and I had been laid off for three months, and I didn't realize the impact the language choice would have.
This was on a medical instrument, so maybe people thought of it as a "safe" language - one that couldn't buffer overflow and corrupt memory. But doing an embedded system with memory-mapped I/O in a Pascal without the Turbo extensions... yeah, that was a less than ideal choice.
