Florian Pflug <fgp@phlo.org> writes:
> In addition, I think we should reword the explanation in 4.2.8 (The SQL Language
> / SQL Syntax / Value Expressions / Window Functions). Instead of that rather
> long (and IMHO hard to read) paragraph about possible frame clauses and their
> behaviour in the presence or absence of an ORDER BY clause, we should go with
> a more algorithmic explanation I think.
Hm. I do not think what you're proposing is a substitute for the text
that's now in 4.2.8, because that's mostly providing syntactic details.
It might make a good addition though.
> If you think that something along these lines would be an improvement, I can try
> to come up with a patch.
Sure.
regards, tom lane