Development system, IDE, toolchain and libraries, as well being better than average in a certain thing as attracting factor.
I agree with that.
Having standalone compilers doesn't mean it is not RAD dependent. At least not any more than C++ does.
It looks like I didn't make the point clearly. Here is the point: if Delphi and Lazarus (the environments) "magically" vanished, the use of Pascal would drop significantly. That is not the case with C, FORTRAN and/or COBOL. Programmers that use those languages don't choose those languages because there is some "cushy" environment available for them, they will probably choose a particular _implementation_ of the language because there is a more flexible/powerful environment for them but, the choice of language isn't determined by that.
For C, Fortran and/or COBOL, as long as there is a decent debugger and editor available in addition to a decent compiler, that's all it takes. For Pascal, if you take away Delphi and or Lazarus, the result will often be "Goodbye Pascal" (very unfortunately.)
If two out of three of (LLVM,GCC,VS) die, you have probably also a significant drop. Even more so if the rumours are true that VS will change to LLVM. (which I still find a bit odd)
Cutting out the most major players always hurts. For any language.
If two out of the three major implementations of the C language were to disappear, the most likely result is that a replacement implementation would appear shortly thereafter. That cannot be said about Pascal.
Quite a few C compilers have come and gone before we got to the current status quo but, C remained popular.
Those languages stood or fell by their 4GL database product. Lazarus/Delphi is different that it makes apps, and a much wider array.
But Pascal has the same weakness those 4GL database products have/had. The popularity of Pascal is dependent on the RAD environments. Only a minority choose Pascal because of the language's power, the majority choose it because of the RAD environments associated with it.
If C had lost Unix in say 1990, most of us couldn't even name it by now.
That's quite likely to be true but, the fact is, it was used to develop a successful O/S. It proved that, in the hands of a reasonably knowledgeable programmer it is powerful enough to write an O/S with it. Pascal doesn't have that and that makes a big difference to programmers that code low level routines.
Cobol is already more or less in that space (except the much regurgitated "in financial institutions", but most of that is heresay rather than actual practical knowledge).
Poor COBOL language. Non COBOL programmers love to put it down. I just want to point out that just because it is hearsay to you, it doesn't mean it is actually hearsay. It is a _fact_, not hearsay, that you'll find a lot more COBOL code in financial institutions than you'll find Pascal code.
but loosing slowly to C++ libraries with scripting languages on top of them.
It's quite likely there is some of that but, I'm rather skeptical that they'll ever be able to displace FORTRAN. You can use a shoe as a hammer but, shoes are very unlikely to displace hammers.
Fortran and Cobol are mostly already there, confined to certain institutions and much less relevant to every day programming than e.g. Pascal.
With the advent of personal computers, "everyday programming" grew significantly. FORTRAN is still mostly used by engineers and scientists and, the number of engineers and scientists didn't grow exponentially with the availability of PCs. The sea got bigger, the number of FORTRAN fish stayed fairly constant while the number of people who really just dabble in programming grew significantly. The result is, the number of FORTRAN fish isn't as noticeable as it once was but, their relevance has not changed significantly.
But the delusion is that it has anything to do with features or planning. They just hit the jackpot somewhere during their existence in being tied to an application that was so big it became self-fulfilling, and which will take a long time to atrophy naturally.
Undoubtedly there is likely a good bit of that in the case of C but, very little of that in the continued use of FORTRAN and COBOL.
Linking it to properties of the language, or some grand plan is just delusional. Chance and other non technical, non-language reasons. Nothing more.
It's not as delusional as you are implying. C is successful in system programming because it demonstrated that it can be used for that purpose _and_ no comprehensive better alternative has showed up. C won the systems programming race in spite of its being a race horse in a wheelchair, because in spite of its wheelchair having dodecagonal wheels, the other horse's wheelchairs had square wheels. Rather regrettably, programmers get used to dodecagonal wheels much too easily (and no matter how poorly they perform, there is always a programmer ready to point out "it works!", therefore no better way should be implemented - that should sound familiar to you.)
That's how it works with other languages too.
But you seem to have forgotten one thing, that is, MS alone cannot let C lose too much popularity simply because it has a multi-billion dollar investment in C code and, they are far from the only company with a significant financial investment in the language.
When it comes to Pascal, Embarcadero may very well be the company with the largest investment in Pascal code and it is paltry compared to the investment other companies have in C code.
It's a lot easier and less costly to abandon Pascal than C (or COBOL or FORTRAN.)
Reality is that it is simply cowardise, and "nobody got fired for chosing X" or "1000000 lemmings can't be wrong" mentality.
It's not cowardice and, there is some, at least at the beginning of a choice, of the "nobody got fired for chosing X" or "1000000 lemmings can't be wrong" but, once a company is heavily invested in a technology, it is often more cost effective to stick with it than adopt a new one even if better simply because it is financially sensible to get the most out of the already built infrastructure.
Making a real decision takes guts,
It sure does. It also takes knowledge and good analytical skills to accurately evaluate the effects of the decision. The latter is what implies the guts because it is too often missing and simply replaced by guts. IOW, the problem is, among the many well developed mathematical fields, there is still no functional "theory of guts".
I sincerely believe that if Pascal was extended to match C feature for feature _and_ it produced code as well (or better) optimized than C's (that's not easy but, it is doable), a very significant percentage of programmers would switch.
Totally ridiculous in my opinion. C's popularity is not about C but about clout. In people minds and in toolchains. You don't solve that with a few haphazardly ported language features.
That seems to be the most common viewpoint but, there was a time when C had little to no clout yet, it grew without it. My premise is, it grew because no really good alternative to it showed up. It seems to be a common hobby these days for some programmers to make another C omelet (rust, D, C++, etc) and proclaim it's better because it is sprinkled with various "safe-whathaveyous" and "multi-this or that". Tunnel vision at its best and, it somewhat works because of the common aversion to unfamiliar constructs.
The bottom line is, the reason there hasn't been a new Pascal standard for so long is simply because the language is not interesting to a sufficiently large audience.