Re: Prolog; was Re: Actions 2022-05-17

John Lumley writes:

> Are we going to mandate that the version is a number/ or a
> multiply-dotted digit sequence ?

That would simplify some kinds of later uses.

But it seems to assume that version identifiers will naturally fall, or
can be induced to fall, into a sequence.  

If we imagine the community group preparing for a choice between doing
things one way or another way -- call them the blue and green approaches
-- by producing variant grammars for comparison, so people can work with
them and it's clear from the labeling which approach a particular sample
grammar assumes, then I think it might be simpler if they had names like
'1.1 blue' and '1.1 green'.  Forcing them both to be numeric imposes a
notion of sequence on them that is at best confusing.  (The history of
HTML version numbers has, if I remember correctly, at least one example
of such a confusion and possibly more.)

I think it would probably be a good idea to decide that we as a group
will use only decimal numbers in published specs while allowing drafts
in progress to have other names, as above.

But I am not sure anyone ever gets this kind of forward-looking
infrastructure completely right, and I am willing to be taught a better
approach.  For the moment, my view is: no one gets it completely right,
but getting it even partly right is worth a lot, and having version
labels is a good step.

Michael

-- 
C. M. Sperberg-McQueen
Black Mesa Technologies LLC
http://blackmesatech.com

Received on Tuesday, 24 May 2022 16:52:40 UTC