Hacker News Reader

Back-compat can have many meanings - but my understanding is that every major release does break some back-compat in some way - otherwise it would not be a major release. And if you think that binary compatibility is not be fair - then how about pseudo-hashes?

This is not to say that I don't agree that Perl has perhaps the best back-compat of all the languages that still evolve - but these absolute statements have this slightly sect feeling.

View on HN

I think pseudo-hashes were always marked as experimental, up until they were removed.

That said, I'm pretty sure you're right that backwards Per 5 broke backwards compatibility with Perl 4 (and presumably 4 with 3, etc.)

From what I've read various places (but can't find a reference at the moment), the version bump from 3 to 4 was just to be able to have the Camel book coincide with a major version (i.e. "this book describes Perl 4" instead of "this book describes Perl 3.1.4.1.5" or whatever).

Reference: http://perldoc.perl.org/perlhist.html

Heh, thanks! I read through the whole thing very carefully before posting, and couldn't find it. Now I see it's in the Introduction: "Perl 4 introduced the first Camel book. Really. We mostly just switched version numbers so the book could refer to 4.000."

Reply on news.ycombinator.com