Re: [gentoo-dev] New project: Gentoo Seeds

[Daniel Ostrow]

On Wed, 2006-09-20 at 18:42 -0400, Alec Warner wrote:
> This whole thread is quite disappointing to me.  Someone comes up with a
> new way to use Gentoo; to make it a viable tool for a job; to make it
> USEFUL.  This is what we are about here (or were?).
>
> "Put another way, the Gentoo philosophy is to create better tools."
>
> 					-Daniel Robbins
> 					Previous Chief Architect
>
> So unless that has changed and no one has updated the webpages...

Here is my take on the issue, it's something I saw happen when Gentoo on
Mac OSX was announced, again with Sunrise, and now with Seeds (also note
I'm not making a value judgment about any of the aforementioned
projects, I just note a similar progression of events). There are those
among us (myself often included, and mostly because I had a hand in the
way the OSX port was handled at the outset) that believe that you
shouldn't announce things in the manner of "Gentoo is doing XYZ now." in
public fora (lists, gwn whataveyou) without first talking internally to
verify the viability of the project, it's impacts on other projects,
potential points of collaboration etc. This also coming up with a
rational reference implementation and a list of tools that you will
need. Now I realize that this means that there is less public visibility
for projects in their larval stage, which can mean less (new) hands
helping to figure out the above, but it also means an informed set of
peers and no surprises.

I believe that what Ciaran (and others) have been trying to say with
suggesting that a GLEP might have been worthwhile isn't so much the
statement that this (or any of the other projects) necessarily *need* a
GLEP per se, but the GLEP process itself can act as a method to hash out
any issues *and* inform your peers. Maybe we just need something along
the lines of a GLPP (Gentoo Linux Project Proposal) mechanism wherein
the Council specifically does *not* need to approve the project, or for
that matter be involved at all, but can, at their discretion, deny the
project existence. The format of the proposal could follow that of the
current GLEP structure, and it's entire purpose would be to foster peer
review and to spread information. Once a general level of consensus, and
not I specifically did not say a full consensus, is reached then the
project can officially be "born".

Hell we just recently went through the whole process of coming up with a
good GLEP to disseminate news to our users and it seems that we have the
same problem internally...

A lot of it comes down to wording in my mind, and granted it is a bunch
of semantic bull but words matter. For instance in Stuart's original
e-mail (and I'm sorry to pick on you, just happens to be the topic at
hand) the subject was "New project: Gentoo Seeds"  and the first
paragraph read "I've created a new project, called Gentoo Seeds [1].
The aim of the project is to create stage4 tarballs which can be used to
'seed' new boxes with ready-built Gentoo solutions." A simple change to
Subject: "New Project Proposal: Gentoo Seeds" with the first paragraph
being "I'd like to create a new project, called Gentoo Seeds [1].  The
aim of the project would be to create stage4 tarballs which can be used
to 'seed' new boxes with ready-built Gentoo solutions. If you are
interested in working on this type of project come by #foo or discuss it
here. I will be sending all online discussions to the list so that the
community can stay informed. Once we get a finalized plan we'll create
an official project." It really comes down to understanding that once it
is called a project it should already be known to be a good idea, and
the whole community should have had time to think about it.

In the court of public opinion there is a huge difference between saying
"Gentoo has a project providing XYZ service." and "Gentoo is looking
into the viability of providing XYZ service." Especially when it comes
to the potential failure of that service. It looks *way* better to say
"We found out that the project would not have been viable." or "We had
to modify our idea in this way to make it viable." then causing what
happened today. I'd also say that the *first* discussion of any new
projects should happen on internal lists with the *first* round of
comments coming from within the dev ranks. That way, if a project is
particularly untenable mention of it won't ever have to be made public.
If it is clear that the project just needs some shake out time then
discussion could move to a public list for further scrutiny and
community involvement.

Again...all semantics...and a load of bull...but bull matters.

--Dan