Re: Clean up glibc version numbers in manual

["Carlos O'Donell" <carlos at systemhalted dot org>]

On Thu, Feb 23, 2012 at 8:00 PM, Roland McGrath <roland@hack.frob.com> wrote:
> I'm not going to get into discussions of the past.
>
>> Anyway, what do people think about that proposal - eliminating UPDATED and
>> generating VERSION automatically? ?(On the basis that a concrete patch
>> would then be posted for people to review the makefile changes involved.)

The purpose of static UPDATED and VERSION is to indicate to the user
when the recipe was last updated, giving them a sense of "Will it apply
to the source base I have?" or "Will it work?"

In response to that statement I say "So what?"

The important cases are:
- Somebody updated it, and it works.
- Nobody updated it, and it doesn't work.

I've been writing recipes for *years* and have never found these
markers useful. Either it works or it doesn't. If it doesn't I
work towards fixing it and submitting a patch or ask for help.
Notice that the UPDATED marker wasn't any help here, what was useful
is knowing if *others* had tried it recently and had it work.

Thus what is *much* more useful is a living document like on a wiki
with an easy feedback mechanism that novice users can use to
indicate "Works" or "Doesn't work."

Developers are head-down trying to fix real bugs instead of testing
recipes and marking the manual with UPDATED/VERSION entries.

My general feeling is that we need to move into this century with
more "living" documentation, but that's an entire project on it's own.

The wiki is the real first step and we're doing a great job there.

I hope my rambling rant made sense.

> That seems potentially reasonable on its face. ?But I think we first need
> to figure out what to do about EDITION and ISBN and how it all relates to
> the FSF's printed editions. ?I haven't talked to anyone there about printed
> manuals since last millenium. ?They appear to be selling a two-volume
> version and I have no idea how the volume split works. ?I think we should
> figure all that out.
>
> If we can't figure it out very soon, then in the interim we should probably
> adopt your suggestion and also remove or (@comment out) EDITION and ISBN
> and their uses, along with anything else that mentions printed editions.

Given your interest in the topic why don't you poke around?

Find out what they need?

Cheers,
Carlos.