Re: Link to bug webpage

Lamar Owen <lamar(dot)owen(at)wgcr(dot)org> writes:
> Let's look at the guy's bulleted list.

> The first item he can't stand is that you can't add a column after any
> arbitrary column, that it goes at the end. Well, this is really
> clueless, as you order the columns when you SELECT or when the
> application presents the data.

Well, I can see some value in it --- but not enough to justify the
implementation pain. It certainly is pretty weak as a leadoff gripe.

> The second item, however, has some real meat in it.

Agreed, we need better ALTER capability. As you say, it's on the TODO
list.

> That third item, about int8. Can a clueless newbie who's heard that
> PostgreSQL is so great, knowing NOTHING about it, find things
> reasonably well in the docs?

He apparently didn't get as far as looking at Table 3-1, on the first
page of the user's guide chapter on datatypes. Still, improving the
docs is an ever-important task.

> However, if it weren't too
> difficult to support index creation at table creation time, why NOT allow
> that? Do we just not _want_ to do it?

We do support it, for UNIQUE indexes (see UNIQUE and PRIMARY KEY
constraints). As for why not plain indexes too, the main answer is that
UNIQUE constraints are SQL92 and any syntax to create indexes otherwise
is not. Of course a CREATE INDEX command is not to be found in SQL92
either, but on the whole I agree with you; this is hard to read as
anything except MySQL's-way-is-the-only-way partisanship.

There hasn't been a lot of talk recently about adopting MySQL-isms, at
least not anywhere near as much as about adopting Oracle-isms. I'd tend
to treat either sort of proposal with suspicion, but we ought to be open
to the idea if we are interested in attracting users of other DBMSs.
Real question is, who out there is excited enough about this point to do
the work?

> Of course, the guy didn't ask on the lists to have it put in TODO. But how
> would he know to ask to have something put in TODO?

I see no evidence that this guy wants to learn about or contribute to
Postgres development at all; he's just looking for things to rag on.
(And not even doing very well at that --- I could name ten worse
problems than these without taking a breath...) The TODO list is
mentioned prominently on the website, for example.

> The last worthwhile item on this guy's list is changing ownership of a
> database. Well, I haven't yet had to do this: can we do this easily?

It could be better. See recent "Multiple Servers" thread over in
pg-admin, notably
http://fts.postgresql.org/db/mw/msg.html?mid=1031042
(which the FTS server seems not to have linked into the thread for some
reason)

regards, tom lane