Re: Process

[Joakim Ziegler <joakim helixcode com>]

On Thu, Nov 16, 2000 at 04:04:22PM -0800, Kyle Olsen wrote:

> We need to follow a development process, this on, or any other that has been
> proven to work. Discussing technologies before the requirements have been
> established is a ridiculous and broken way of doing things. A clear project
> leader is needed as well, to organize and guide us. I assume that's going to 
> be Joakim. :)

Process indeed has a place. However, knowing the GNOME project, it's probably
going to be somewhat less structure-heavy than your typical commercial web
development project. This is nothing to be afraid of, although it can be a
unfamiliar thing at first. :)

Basically, if people want to talk about technologies to use, I'm not going to
prevent them. There are a couple of reasons for this, first of all that
mostly all the technologies that have been discussed have the same basic
capabilities. Having done a large amount of sites, and having quite a bit of
experience on GNOME sites, I can't think of any issues that would come up in
a requirement specification that would seriously impact the choice of tools.
Python, Perl, PHP, all have the same basic capabilites. (Also, of course,
it's not like we have zero idea of what we're building. I'm not expecting the
new sites to be horrifyingly different in content and structure from the old
ones, in particular since the content and structure on the existing
www.gnome.org is the result of a very deliberate requirement specification
process, although the implementation is lacking in places).

Also, what's worth noting is that the very volunteer basis of this project
makes discussing the tools early on important. I'm *extremely* concerned that
decisions should be made by the people who are actually going to be working
on the project, to make it as comfortable as possible, and I know there are
people who have strong feelings about what tools and technologies to use,
perhaps strong enough that they'll not want to work on the project if a
specific technology is chosen or not chosen. Thus, it's good to have this
worked out at least roughly to begin with, so we know who is going to be the
team. This is a pretty unique thing about volunteer free software work like
this, since if it was a commercial project, engineers would probably just
toughen up and do the work anyway, even if there were decisions made that
they don't agree with. So the dynamic is quite different.

So, in general, I think the process you're outlining is good, but a little
formal and a little serial to fit 100% to this project. A degree of
parallelism is good, and I think having a technical discussion alongside a
high-level requirements discussion doesn't hurt.

Additionally, there's now a thread that actually calls for people to
contribute to creating a requirement specification, and that's exactly where
people like you, who have experience building a lot of sites, should be.

Welcome on board.

-- 
Joakim Ziegler - Helix Code web monkey - joakim helixcode com - Radagast IRC
      FIX sysop - free software coder - FIDEL & Conglomerate developer
            http://www.avmaria.com/ - http://www.helixcode.com/