Joel Spolsky is a troll. He occasionally says something of value, but
then, a stopped clock is right twice a day.
Joel might be a troll, but I've gotten a lot out of his writings. I
think they are pretty good, and I like to go to his site and read his
essays (On Software) from time to time.
I agree that Perl has been mostly forgotten by the world at large, but
I don't think that's the important point. I've had two experiences in
the last four weeks, introducing Perl (or rather, Perl apps) to IT
guys who came into the conversation expressing a lot of skepticism
about Perl. However, after looking at the code and comparing it head
to head in one case with the identical app written in Java, they had
to admit that Perl was a legitimate tool for some jobs.
One poster here (I think it was Uri but I might be mistaken) recently
observed that Perl was still the best at its original purpose,
extracting and reporting data. If many have forgotten Perl or never
knew it in the first place, it's because they apparently don't have a
need to extract or report data. This isn't a criticism, merely a
statement that people don't normally use tools that they don't need.
I read this week a report on using Erlang as a web development
language. I am not ashamed to state that I use Perl for web
development and have experimented a number of different languages such
as Python, CFML, .NET, Java, and others. The report had a section
explaining why Erlang was appropriate for web apps which was pretty
persuasive -- multithreading, stability, hot fixes, and so on. The
only problem is that people haven't used Erlang for web apps, but they
have used a lot of Perl. Over a period of time people will tend to use
the tools best suited for particular jobs, and that includes Perl, but
not Erlang. I have a pin extractor which I use rarely, but it's
absolutely indispensable when I need to extract a pin, and in the same
vein, I wouldn't pay much attention to reports that people have
forgotten Perl.
CC