=?ISO-8859-1?Q?Dennis_Bj=F6rklund?= <db@zigo.dhs.org> writes:
> I guess it's specified in the SQL spec, but how come a tuple is not an
> expression? If it had been an expression the first part of that section
> would still apply, which is why I just read the first part.
Well, if you want to think of "(42, 'foo')" as being an expression then
I suppose it could be considered to be all one syntax. Personally I
think that'd be more confusing rather than less so.
Another alternative is to put both syntax summaries at the top of the
subsection, and combine the two textual descriptions into one.
This is the wrong place for this discussion, though. If you want to
have a go at rewriting the section, why not put up a sketch on
pgsql-docs and see if people like it?
regards, tom lane