>>>>> "AS" == Andreas Schott <[EMAIL PROTECTED]> writes:

AS> So what is actually the latest official release? I strongly
AS> recommend releasing at least two offical versions per year.
AS> But this also means, since we now have 1999, if you release
AS> v98.2 beta8 you should name it v99.1.

Attached is a msg containing Marek's response to a similar plea I made
about a year ago.

Before reading it, please understand that the guys working on L2H are
my heroes, esp. Ross, who has never failed to rip into any problems
I've run into.

-- 
John A. Turner, Ph.D.                Senior Research Associate
Blue Sky Studios                     http://www.blueskystudios.com/
One South Road, Harrison, NY 10528   http://www.lanl.gov/home/turner/
Phone: 914-381-8400                  http://john.turner.org/

Info on Blue Sky's fully computer-generated short film, nominated for
an Academy Award this year:
    http://bunny.blueskystudios.com/




On Mon, 2 Mar 1998, John Turner wrote:

> Ross Moore writes (in part):
> 
>  >  in v98.1p1 (patch-level 1, to be released ASAP)
> 
> A humble request to the developers.  Use the minor version number.
> 
> Granted, it may be a bit soon for 98.2, and this patch may indeed be
> too minor to warrant an increment in the minor version number, but
> I'd like to request more frequent "official" releases.  We had 96.1
> and a zillion patches, 97.1 and a confusion of developmental
> releases, and now the outstanding 98.1.

You are probably right. I really see no reason why we should not increment
the minor version number. Even if we arrive at 98.56 at christmas ;-)

> I don't mean to be an ass, and I hope this doesn't come across as
> whining, but I think it would be beneficial to the community.  When
> there's an important fix, go ahead and bump the minor version number
> and release.  In the release notes mention what was fixed and people
> can decide for themselves whether to upgrade or not.

Right. A release is a release in the sense that it appears on the official
distribution site. And it is easier to deal with major.minor than
major.minor.patchlevel.

On Mon, 2 Mar 1998, Fred L. Drake wrote:

> John Turner writes:
>  > A humble request to the developers.  Use the minor version number.
> 
>   I second this request.  It seems that versions are released and
> then bugs get fixed, and a similar version replaces the original with
> no change in the version number.  Perhaps this only happens in the
> short period immediately after a release, but it can make it hard to
> know just what you have or to tell somebody else "use release
> Foo.Bar".  I don't mind having a patchlevel in addition to the
> major.minor numbers, but what would be most helpful with versioning is 
> that once code is released, the code that matches that version number
> should never change.  A new patchlevel really is a new release!

100% right. AFAIK nobody really changed an existing release. This does of
course not pertain for the development container at Darmstadt.

I do not want to start a war about versioning schemata. I am really very
happy with the existing format. For the developers there is still the CVS
version that exactly indicates the version of each file. I'll add some
code to display this version when using the -help or -version switch.
My suggestion is to increase the minor version number in each official
release that appears at the official latex2html site, although this
approach might conceal the "code quality", i.e. the maturity of the code.

Whew, perhaps I should think about it one more time. With the upcoming
changes in (what was formerly known as) 98.2 ;-) in mind, I'd rather
suggest we should stick at what Ross says:

98.1    : official release one in 1998
98.1p1  : first patch release. Only bugs fixed
98.1p2  : some more bugs fixed
98.2    : second release in 98, with enhanced features compared to 98.1
          (*and* all the patches of 98.1)
98.2p1  : some bugs in 98.2 fixed

Note that each version should contain the *complete* release, not only the
patches, so that it's easy for a user to get a certain version.

Maybe we should move to a three-digit-scheme like the Linux kernel has?

Jens, Ross, Daniel, what are your opinions?

Cheers,

Marek

***********************************************************
* Marek Rouchal         [EMAIL PROTECTED] *
*               http://saftsack.fs.uni-bayreuth.de/~marek *
* Linux, Perl, Latex2HTML enthusiast. PGP key available.  * 
***********************************************************



Reply via email to