On Wed, 30 Apr 2008, Robert Treat wrote:
> Whenever anyone posts a problem on 7.3, the first thing people do now
> days is jump up and down waving thier arms about while exclaiming how
> quickly they should upgrade. While I am certain there are even older
> versions of postgres still running in production out there, I'd have to
> say that the core developers for this project do not release software
> with the expectation that you will use if for more than 5 years.
You could easily make a case that 7.3 wasn't quite mature enough overall
to be useful for 5 years. There's little reason to keep pumping support
effort into something with unfixable flaws. I know when I was using 7.4
heavily, I never felt like that was something I could keep going for that
long; the VACUUM issues in particular really stuck out as something I
wouldn't be likely to handle on future hardware having larger databases.
8.1, on the other hand, is the first release I thought you could base a
long-term effort on, and 8.2 and 8.3 have moved further in that direction.
8.1 has been out for 2.5 years now, and it seems like it's got plenty of
useful left in it still (except on Windows). The improvements in 8.2 and
8.3 are significant but not hugely important unless you're suffering
performance issues.
Compare with 7.3, which came out at the end of 2002. By 2.5 years after
that, the project was well into 8.0, which was clearly a huge leap.
PITR, tablespaces, whole new buffer strategy, these are really fundamental
and compelling rather than the more incremental improvements coming out
nowadays.
(Obligatory Oracle comparison: for customers with standard support
levels, Oracle 8.1 was EOL'd after slightly more than 4 years. It wasn't
until V9 that they pushed that to 5 years)
--
* Greg Smith gsmith@gregsmith.com http://www.gregsmith.com Baltimore, MD