On Mon, Apr 27, 2009 at 10:56:28PM +0300, Peter Eisentraut wrote:
> On Monday 27 April 2009 21:54:12 David Fetter wrote:
> > On Mon, Apr 27, 2009 at 09:50:13PM +0300, Peter Eisentraut wrote:
> > > On Monday 27 April 2009 20:10:27 David Fetter wrote:
> > > > While we're at it, can we see about moving foreign keys out of
> > > > the "advanced" section? They've been standard for many years.
> > > > Heck, even MySQL has had them, at least in some of their
> > > > engines, for many years.
> > >
> > > Advanced doesn't have to mean nonstandard, and standard doesn't
> > > have to mean basic.
> >
> > Are you seriously arguing that foreign keys aren't basic?
> > Seriously?
>
> In your words: yes.
>
> But it's all relative. Among all the topics that are covered in the
> tutorial, foreign keys have certain prerequisite topics, such as
> logging into the database, creating tables, and putting data in.
> Certain things have to come before others, and sections are used to
> organize the information. You can relabel the sections to "Really
> Basic" and "Somewhat Basic", if it helps you.
How about "setting up," and "usage," which opens the door to more
advanced features. In later revs, we could have sections with "Good
Practices" in the title.
> In some way, everything that is covered in the tutorial ought to be
> "basic". But some things are more basic than others. You could
> equally make the argument that views and transactions are basic, but
> then there would be hardly anything interesting left in the
> "advanced" section,
Actually, there's plenty.
> especially if you add the argument
*I* certainly wouldn't add that argument.
> that inheritance and window functions could be considered by some as
> very esoteric features that shouldn't be in the tutorial at all.
Partitioning isn't exactly an esoteric feature. More of an advanced
topic, as windowing and OLAP functions are, at least this year.
SQL/MED might qualify as esoteric, at least this year. Five years
hence, SQL/MED might be pretty basic, as most of the heavy lifting
will likely be in the back-end, at least for the common scenarios. By
then, the optimizer will have what it needs to do what it does fairly
transparently. :)
My point in all this is that, "advanced" isn't frozen forever in time,
and neither is, "basic."
Cheers,
David.
--
David Fetter <david@fetter.org> http://fetter.org/
Phone: +1 415 235 3778 AIM: dfetter666 Yahoo!: dfetter
Skype: davidfetter XMPP: david.fetter@gmail.com
Remember to vote!
Consider donating to Postgres: http://www.postgresql.org/about/donate
