Thread: Section 9.7.3.2. Bracket Expressions should be rewritten

The following documentation comment has been logged on the website:

Page: https://www.postgresql.org/docs/14/functions-matching.html
Description:

I am a Postgres newbie, and I find this part of what is generally excellent
documentation quite confusing. It contains the following disclaimer:

Note
PostgreSQL currently does not support multi-character collating elements.
This information describes possible future behavior.

I am very skeptical of the idea of documentation describing "possible future
behavior." Unless carefully done, this is guaranteed to cause confusion to
newbies such as myself. In this case, it is very unclear as to what the
disclaimer refers to. Is it referring to the previous 3 paragraphs, the
following 3 paragraphs, some subset of those paragraphs, or the entire
section? If reference to possible future behavior is deemed absolutely
necessary, then IMHO that portion of the docs should be blockquoted and
clearly marked as being forward-looking and not part of present behavior. 

Furthermore, it appears that this disclaimer has been part of the docs since
at least version 10, so it's not even clear that this feature is even
currently being seriously considered for implementation.

PG Doc comments form <noreply@postgresql.org> writes:
> I am a Postgres newbie, and I find this part of what is generally excellent
> documentation quite confusing. It contains the following disclaimer:

> Note
> PostgreSQL currently does not support multi-character collating elements.
> This information describes possible future behavior.

> I am very skeptical of the idea of documentation describing "possible future
> behavior."

What that means is "we don't implement this, but it's required by POSIX,
so we really ought to get around to it someday".

I have considered the idea of deleting mention of multi-character
collating elements from the preceding paragraph, but then we'd be
left with nothing but a statement that "[.x.]" means the same
thing as "x" inside a bracket expression.  Which is weird and
confusing enough that we'd still have to talk about multi-character
collating elements in order to explain why it's like that.

> Furthermore, it appears that this disclaimer has been part of the docs since
> at least version 10, so it's not even clear that this feature is even
> currently being seriously considered for implementation.

Hah!  Look a bit further back :-(

            regards, tom lane