On Mon, 2004-08-02 at 19:57, Steve Bergman wrote:
> On Mon, 2004-08-02 at 16:47 -0700, Jonathan Gardner wrote:
>
> >
> > They don't need to run postgreSQL on their developer machine. Just provide
> > them with a test cluster or a developer's cluster and allow them to create
> > and administer their own databases.
>
> At work, I do run a MySQL and now also a PostgreSQL server. However,
> one of my coworkers is taken in enough with the idea of using Open
> Source tools that he wants to set up a development platform at home.
> Historically, a Windows developer, he is installing Linux there.
> However, he wants the ability to run the same stuff under Windows for
> our clients who don't use Unix or Linux.
Good news, PostgreSQL is about to enter beta for the next version, and
will include a native windows port as part of that version. Some other
cool features will be Point in Time Recovery (allows restoring a
database to EXACTLY the point it was at when the storage subsystem it
was on dies,) Sub transactions / save points, and a few other nice ones,
take a look at the release notes when beta comes out this week or so.
> I really appreciate the great responses that I have gotten here. I have
> ported my (rather small and simple, 1500 line, 4 tables) current project
> to pgsql, and have already used a few pgsql features. Namely, the (more
> flexible) unique key constraints and also views. When I decided to
> develop the pgsql and mysql branches in tandem, I decided to guard
> against trying to use pgsql as a mysql drop in replacement, since I knew
> that would be unfair to pgsql. I've decided, at this point do drop
> development on the mysql branch and just finish up the pgsql branch.
> I'm pretty much convinced that learning pgsql is where I should put my
> efforts. Once I've had some more time to study pgsql's capabilities,
> I'll probably tear this project down and completely redesign it to take
> full advantage.
That was pretty much my philosophy back when I started with version
6.5.3 of PostgreSQL back in the day. Back then it was kinda slow and
you had to be a little more careful what you were doing, as postgres was
still a little fragile. Things like unconstrained joins could kill the
whole server instance. Nowadays it's very robust, and quite fast. And
every time I turn around someone's added some useful new feature I
didn't really know I needed.
