[Date Prev][Date Next][Thread Prev][Thread Next][Author Index][Date Index][Thread Index]
Re: [zigzag] GZigZag 0.5.0 has been released]
- To: ZZ Development <zzdev@xxxxxxxxxx>
- Subject: Re: [zigzag] GZigZag 0.5.0 has been released]
- From: Antti-Juhani Kaijanaho <gaia@xxxxxx>
- Date: Mon, 22 Jan 2001 21:00:06 +0200
- In-reply-to: <firstname.lastname@example.org>; from eharter@xxxxxxxxx on Tue, Jan 23, 2001 at 03:19:04AM +0900
- Mail-followup-to: ZZ Development <zzdev@xxxxxxxxxx>
- References: <email@example.com>
- Sender: Antti-Juhani Kaijanaho <ajk@xxxxxxxxxxx>
On 20010123T031904+0900, Edward Harter wrote:
> Fine; stable is a long-established term and has a _number_ of uses in
> software development, and in other areas of development.
True. I am referring to its meaning in release management, as I have
seen it used.
I have mostly observed and participated in one huge free software
project, namely Debian GNU/Linux (currently over 500 volunteer developers
and over two gigabytes worth of compressed executable code for IA32).
The only accurate definition of "stable" I have seen used there is indeed,
"not moving". Meaning that we know that there are bugs (some of which
we may be aware of and for some reason have not yet fixed), but since
it is a stable version, you know how to work around them, instead of
having to confront different kind of breakage every day.
I do regard 0.5.0 as the best version of GZigZag available. However,
I do not believe - as a user or as a developer - that your definition
of stable is reasonable. It creates unfounded expectations on the
stable software, for very few software projects really achieve the
level of reliability that you would expect from, say, a modern building.
Most of those that do - systems whose failure would cause a disaster,
such as space mission control systems or nuclear missile control systems
- are built under such requirements that would make more experimental
software such as GZigZag practically infeasible in the short run.
When we have considerable experience with the prototype, we may well be
able to produce a reliable system. But currently the prototype is too
volatile for us to stabilize it that well.
> Is there something in what I am writing that isn't sufficiently clear
> to you AJ?
> Or do you feel that these desires on the part of the user are
> unreasonable? Can that possibly be so??
Please don't be offended. My primary goal is to make GZigZag better,
I do release management on the side. And my views on release management
may be a little fatalistic, but I genuinely think that whatever the
current stable version is, it is the best version to be doing things with.
Note that I personally don't consider GZigZag to be ready for real work
yet, in any existing version. It does pretty and amazing demos, and it
has promise. But I don't use it for anything significant yet.
%%% Antti-Juhani Kaijanaho % gaia@xxxxxx % http://www.iki.fi/gaia/ %%%
Keep the Deja Archive Alive!