Re: wiki [Was Re: The trick to build PySablot and an updated SuSE-GNUe-HOWTO]

[Peter Sullivan]

On Sat, Oct 19, 2002 at 12:08:03PM -0700, Derek Neighbors wrote:
> On Sat, 2002-10-19 at 10:07, Jason Cater wrote:
> > I know that some don't like the idea of a GNUe wiki, but I'd like us to
> > consider it again. Someone mentioned they didn't like the idea of doing
> > docs in wiki at the expense of some traditional format. 
[...]
> Maybe our web site just needs an easier
> way to upload 'user contributed' documentation?

/me waits for someone to suggest phpNuke/postnuke as a possible solution...
I like the concepts behind wiki, and I think it would fit well with 
both what we are now and what we want to become as a GNUe community. My 
concern would be that we already have multiple channels of communication 
on this project, and I'm not sure that adding yet another one will help. 
We already have:

1. Readmes etc. in CVS
2. Tech Notes in CVS
3. Formal Documentation in CVS (docbook, lyx, html etc.)
4. The 3 current mailing lists, not all of which everyone gets
5. The historic mailing lists (are these still publically accesible anywhere?)
6. The website (improved, but still needs much work)
7. The IRC logs (which could really do with better indexing than "grep")
8. DCL for bug tracking (a bit more specialised, I know, but still 
   a comms channel) 
9. Kernel Cousins (summary of 4. and 7.)  

The net effect of all of these is that not only newcomers, but also the 
core people, can find it difficult to keep in touch with everything that 
is going on (It's both flattering to me but also slightly worrying how 
even the core developers say they like Kernel Cousins as it's the only 
way to keep up). And this is with the project at a point (hopefully) just 
*before* it starts to take off exponentially... This probably links to 
Jason's point in another thread about over-management.

So, I would support a wiki, and even be prepared to help set it up and 
"seed" it, provided that we used it to *replace* some of the above 
channels instead of just letting it become number 10 on the list. It 
could replace 1, 2, 5, much of 6, maybe 8 (although I suspect the 
"eating own dogfood" argument applies here), and become a definitive 
pointer/indexing mechanism for 3, 7 and 9. 

I recognise that this is almost the complete opposite of what other 
people in this thread see a wiki or wiki-style forum for. I also 
recognise that, by writing this e-mail rather than doing some more 
work on the website this evening as intended, I am contributing to 
the over-management that Jason refers to...

-- 
Peter Sullivan <address@hidden>