Re: [bug #62936] Confusing description of chained rules in the manual

[Paul Smith]

On Sat, 2022-10-22 at 21:13 +0100, Tim Murphy wrote:
> On Sat, 22 Oct 2022, 20:33 Paul Smith, <psmith@gnu.org> wrote:
> > "Updating a target" has a well-defined meaning that's used
> > everywhere in the documentation and the mailing lists etc., and
> > it's also the term used everywhere in the POSIX spec for the
> > definition of make and in the documentation of other versions of
> > make such as BSD make etc.
> 
> I didn't suggest upsert but a new word which you can make accepted by
> virtue of gnu make's very important position.

I disagree that using an invented new word is an improvement.

> Manuals surely are primarily written for new users not people who
> know it all already.

Very true, and one of the primary purposes for manuals is to introduce
new users to the important concepts and terminology used with the tool
being documented.

> Anyhow it is easier to do nothing.

Yes, and sometimes it's also better as well.

I've never claimed the manual is perfect, or even all that great.  I'm
happy to reword this section to make it more clear, if someone explains
where the confusion is, but I'm not going to replace "updating" because
that term is central to how make (and not just GNU make) works.

> 
> To me that seems like a tortuous justification to win a debate but
> obviously your word is final.

I guess we're even then, because to me the appeal to Cambridge and "Is
make really "trying not to make things worse?"" seemed like a tortuous
justification for objecting to that phrase.

Cheers!