On Fri, Apr 7, 2017 at 01:06:53PM -0400, Tom Lane wrote:
> Bruce Momjian <bruce@momjian.us> writes:
> > This change is not on my patch yes but what do you think about changing
> > "regular" and "normal" aggregate designations to "generic"? That seems
> > clearer because the aggregates work in non-window and window contexts.
>
> I don't think "generic" is le mot juste here. What's more generic about
> avg() than about mode(), for instance?
>
> My dictionary says "generic" means "characteristic of a class or group
> of things", which doesn't seem to have anything to do with this.
>
> I actually think "regular" might be a better choice than "normal" for this
> purpose, mainly because somebody looking for statistical aggregates might
> think that "normal" has something to do with normal distributions. So
> maybe we should leave that one group of uses alone and s/normal/regular/
> in section 9.20. But in any case it would help to define the term
> explicitly in both sections.
I don't think adding another defined term in this dense text is going to
help so I just specified "general-purpose and statistical aggregates"
where needed; patch attached.
I didn't touch any mention of "normal" in reference to
pg_aggregate.aggfnoid where normal/"n" has a specific meaning.
--
Bruce Momjian <bruce@momjian.us> http://momjian.us
EnterpriseDB http://enterprisedb.com
+ As you are, so once was I. As I am, so you will be. +
+ Ancient Roman grave inscription +