What would it take to convince you to accept a job coding in an "obsolete" language every day, like COBOL, Pascal, or Fortran?
Is it money, and there's a salary number? Is there a unique benefit (perhaps an annual 3 month sabbatical or 20-hour work weeks)? Or do you love an obsolete lang, and are already looking for that job?
COBOL - probably a lot, but not due to the language - it'd be 30 years of accreted bad programming that would suck.
Fortran - depends, like Pascal, on dialect. The most recent ones are quite reasonable.
I'm not sure that Pascal was ever significant enough to become obsolete. It was intended as a teaching language. It got used for a variety of things anyway, but was replaced by better languages before it got much traction.
I'll write COBOL, Fortran, or Pascal -- or Algol or Snobol or Logo. I don't mind legacy systems; I don't often get a chance to clean-sheet things.
The fact that those dead languages are running on either decrepit computers, or emulators for decrepit computers, that is less fun. Ideally, you'd be there to work on a plan to replace it with something more capable, because otherwise your project is permanently hindered by the original hardware.
But meantime, I'm perfectly fine keeping something running just because people need it to keep doing its job. If they can't afford to pay for an upgrade path, that's worrying, but only because they can't afford to pay for anything. Ya gotta spend money to make money.
Languages just don't make much difference to me. Frameworks make a bigger difference -- and yeah, for older languages, framework and language are pretty much the same thing. But all I want from a framework is to get the thing to do its job. The rest is just programming, and I'm happy to program. The fundamental problem-solving skills apply regardless of the language.
While probably unlikely, if someone wanted me to do a new project in Fortran, COBOL, or Pascal it would be easier than if I got thrown into the middle of an existing mess and told to fix it. That can be really challenging, because it's amazing how bad some code is.
On the other hand, if getting paid $175+ an hour, a job is a job. If it takes longer because the code is shitty, well, it also pays more.
Money is money after all
I would be less inclined to take a MUMPS gig than a Forth one. Part of the issue would be the environment. Developing Pascal on a modern unix environment with docker and all the nice tools is one thing. Having to go back to Turbo Pascal on MSDOS 3 would be a different proposition
Some fraction of Fortran programming is still done in FORTRAN 77. I would demand more to program in FORTRAN 77, maybe 100K-200K more, than to program in Fortran 2018.
Also: don't touch ISAM. You've been warned