I disagree. That's not the definition that people use when they say that a human language is dead, and it shouldn't be the definition people use when describing a computer language.
Latin is a dead language. That doesn't mean that no one speaks it, nor does it mean that no literature is available in Latin. It's still an official language of the Vatican. There's also Nuntii Latini, a news service which broadcasts in Latin. They've had to make up vocabulary for concepts which didn't exist 2,000 years ago.
But Latin is still a dead language.
Similarly, PL/I is a dead computer language. Sure, there's still support for it (https://www-01.ibm.com/common/ssi/cgi-bin/ssialias?subtype=c... ). But it's a dead language.
So is SNOBOL and its many variants.
So is TRAC, despite the many fans who learned about it through Nelson's 1974 book "Computer Lib".
Hieroglyphic Egyptian is a dead language, even though people can read and write it.
So is cgi-lib.pl, which was the most popular web framework back when Perl4 ruled the web. (I checked; last update was 19 years ago.)