Re: Wish: remove ancient constructs from Postgres

On Sun, 2006-02-26 at 12:08 -0500, Tom Lane wrote:
> We'd consider removing these features if they were actually blocking
> support of some spec-required behavior ... but since they don't, it's
> quite unlikely they'll ever be removed.

Right; there are plenty of places in which PostgreSQL extends the
standard. If you're concerned about writing standard-compliant
applications, merely removing the places where we have historical syntax
variants is probably going to be of little help.

I think a better approach would be to introduce the concept of "SQL
dialects", similar to "--std=..." in GCC or SQL modes in MySQL 5. That
would help people who want to write standard-compliant applications
while not inconveniencing those who don't care.

-Neil