On Mon, 13 Aug 2007, Bill Moran wrote:
> Another minor nitpick ... I noticed references to both 8.2 and 8.1, and it
> felt slightly inconsistent to me.
I was trying to establish that performance comparisions made against
PostgreSQL versions before 8.1 shouldn't be considered relevant today
because of the scale of improvements in that particular release, but that
either 8.1 or 8.2 are speed-competitive when fairly compared again MySQL 5
using InnoDB+strict mode. Next time I'm editing in there, I'll try to
state that more explicitly.
I don't think it's a good idea to only talk about 8.2 when so many vendors
are still shipping 8.1 as their mainstream supported release. I'd hate
for someone who was considering using PostgreSQL on, say, their shiny new
RHEL 5 box go through this thought process: "oh, that comes with 8.1.
That paper we read only talked about 8.2, so that must be the old slow
version. We shouldn't use that, and since we'll lose support if we don't
run RedHat's package we can't upgrade to 8.2. Scrap that idea, we'll have
to use MySQL if we want it to be fast".
And let me cut off any factual arguments that it doesn't work that way by
saying I have this particular discussion ("we'll lose support if we touch
any of the vendor packages") all the time with people, and it's a very
common perception of how things work among general businesspeople whether
or not it's true for a particular software product.
--
* Greg Smith gsmith@gregsmith.com http://www.gregsmith.com Baltimore, MD