[wp-hackers] WordPress 3

Alex Günsche ag.ml2007 at zirona.com
Fri May 18 12:01:58 GMT 2007


On Fri, 2007-05-18 at 06:55 -0400, Doug Stewart wrote:
> Gentoo and Ubuntu have been doing that for a while and, well, I don't
> like it.  I'd have a hard time putting a finger on it, but it just
> seems, well, weird.  While the releases tied to a year (or, as in
> Madden's case, tied to the NEXT year) make sense for a product
> released once-yearly, it somehow just comes off as a bit geeky for
> twice- or thrice-yearly.  (Feisty Fawn being 2007.4, for instance.)
> 
> The nice thing about a release numbering scheme not tied to chronology
> is that it leaves you free to indicate to users the relative "weight"
> you give a release.  Since we've been talking about what would
> constitute a 2.x->3 jump, I think we're keyed in on this notion fairly
> well.
> 
> People have been using X.Y numbering schemes since almost the
> beginning of computer science and there's a certain expectation built
> into that at this point.  People look at a 3.0 product and figure
> "Hey, must be some big changes from 2.x" whereas there's no way to
> evoke that reaction when they say "Well, 2006.10 was okay, but 2007.4
> is where it's REALLY at!"
> 
> Hope I'm making some modicum of sense -- it's early and I've not yet
> had any caffeine.

I absolutely agree with every single word! I also don't like year-based
versioning, for the same reasons.

For example, when there are big changes during one year, those versions
would be 2007.0 to 2007.23 while the first bugfix in 2008 would be
2008.0 -- that doesn't make sense.

And just imagine Linus would release kernels like 2001.6.20.4... ;)

Alex

-- 
Alex Günsche, Zirona OpenSource-Consulting
http://www.zirona.com/ | Hilfe für das HQ AC: http://www.prohq.de
PubKey for this address: http://www.zirona.com/misc/ag.ml2007.asc



More information about the wp-hackers mailing list