bart wondered why Text::Metaphone jumped to 2.0 for just a bug fix. It prompted me to share my views on version numbers.
There's two important things about version numbers.
1) They sort properly.
2) They always go up.
Assume anything outside of that and all bets are off. This is why I'm considering switching to plain integers.
But people assume lots of things about version numbers, the biggest misnomer is that big version jumps must mean big changes. This prompts people to purchase the upgrade. This is why during the .com boom version numbers became marketing tools. Solaris 2.7 became 7, Java 1.5 became 5 and Perl 5.6 isn't 5.006. Microsoft labeling all their releases by year is pure marketing brilliance. Every time you load up Microsoft Word 2005 there it is, right in your face: "YOUR SHIT IS OLD! GO OUT AT BUY THE SHINY NEW SHIT!" If we called 5.8.0 "Perl 2002" maybe more people would upgrade.
If we called 5.8.0 "Perl 2002" maybe more people would upgrade.
I really like this idea. I'm adding this second sentence so it doesn't trip your one-sentence rule for me; I really do like this idea.
Perl 5.10 is going to really bugger the sortability then, cos it's going to go 5.005, 5.10, 5.6, 5.8. Oops!
Re:Perl 5.10
schwern on 2007-07-02T20:14:21
I didn't say *how* it sorts, just that it sorts "properly". Meaning works within the sorting rules of the system. 5.10.0 is following the perl version sorting rules.
It worked for Subversion!
This is why I'm considering switching to plain integers.
