Re: Fwd: version vs edition numbers in Emacs manuals

[Paul Eggert <eggert@cs.ucla.edu>]

On 11/14/19 4:18 AM, Eli Zaretskii wrote:

> I didn't see John's response, not sure why.  Can someone forward it to
> me?

Sure, here it is:

> Paul Eggert <eggert@cs.ucla.edu> writes:
> 
>> On 11/10/19 6:25 PM, John Sullivan wrote:
>>> Can you fill me in more on how they are messed up? Thanks for your help.
>>
>> For example, the FSF bookstore web page
>> <https://shop.fsf.org/books/gnu-emacs-manual-18th-edition-v-261>
>> currently advertises the GNU Emacs 26.1 manual as the "18th edition,
>> v. 26.1" and the spine of the printed manual says "Eighteenth edition
>> for GNU Emacs version 26.1".
>>
>> In contrast, the online manual
>> <https://www.gnu.org/software/emacs/manual/emacs.html> says on its
>> title page "Seventeenth Edition, Updated for Emacs Version 26.3". That
>> is, the online manual says it is for an *earlier* manual edition, but
>> for a *later* Emacs version, than the printed manual.
>>
>> The edition number is incorrect and misleading, since the online
>> manual is in fact more up-to-date than the printed manual. So I'd like
>> to remove the edition number from the online manual. That way, the
>> online manual won't have incorrect information, and the people
>> printing the book can use whatever edition number they like without
>> having to coordinate with the Emacs developers.
>>
> 
> That makes sense to me. The Emacs developers do prepare the print
> version for us. We might like to keep an online version that matches the
> print version. So maybe the solution could be to have two online
> versions, one that corresponds to print and is only updated when a new
> print version is made, and the other is the one that the developers with
> each release or whenever they feel is appropriate? And that latter
> version would not have any edition number. We wouldn't have any problem
> with the primary canonical manual link
> (https://www.gnu.org/software/emacs/manual/emacs.html) being used for
> the online-only version. What do you think?


Getting back to Eli's email:

> we should have a separate target in the Makefile, and we should maintain
> the last printed edition in some separate file, because no one will
> remember that otherwise.

A separate Makefile target would be fine, but the separate file should 
be something that the FSF Press maintains. Having the edition numbers be 
in a file that Emacs developers maintain would continue to cause 
glitches like the ones noted above. The FSF Press is downstream from 
developers, they generate edition numbers at their convenience not 
developers', and they can and should be the ones who keep track of the 
edition numbers that they maintain.

> The fact that some manuals use EDITION while others use VERSION also
> doesn't make this very clean, IMO.

Yes, that area could easily be made more systematic. For example, we 
could systematically use just EDITION and DATE for all the FSF 
Press-maintained info.