> > > If somebody has a new class or library, then they should add it to
> > > the RAA index.  Want to see what's new? Go to the RAA.
> >
> > NO, WRONG!
> 
> What, specifically, is wrong? Listing items on RAA?  Using RAA as a
> central point of information about contributed libraries?  Expecting
> people to actively check the RAA for changes?

Expecting people to goto the RAA.  (ps, sorry for being harsh/brash)

> > This is the same broken attitude that promotes the use of Wiki's,
> > which, as far as I'm concerned are a black hole for information
> > and should be avoided at all costs.  Want to propagate
> > information?  Setup a cvs commit emailer that sends out diffs.
> > Fire up a docbook project and hand out cvs accounts left and
> > right.  What do you have at the end of a wiki?  A chunk of HTML
> > and a community.
> 
> There's a lot to be said for a chunk of HTML and a community.
> 
> There are whatever number of ways to propagate information.
> Face-to-face, IRC, mailing lists/newsgroups, Wikis, "conventional"
> web sites, books, etc.  I can't see dismissing any one them outright
> as indicative of a broken attitude, though they can be misused.
> Now, at the extremes, I can't see using either IRC or books as a
> primary means of dissemination announcements, but what's suitable
> depends on the shelf life of the information, and whether some
> degree of interactivity is called for.

I don't dismiss it, but I equate wiki's to postit notes that are
chained together and stuck on a bulletin board.  Works for basic
information sharing, but it's far from a tomb of information.  As
dblack said on IRC, redirecting a conversation to a wiki is the kiss
of death for the topic.  Wiki's are seductively easy to use, but are
blackholes, nevermind that you have to go visit the bulletin board to
see what's changed.

> Now, if some people decided to check updates from rubyxml.com (or
> wherever the RSS feed is presented) rather than rubygarden.org
> directly, that would distribute the load.  Push or pull? Centralized
> distribution, or distributed storage?  What are the trade-offs?

Email is an event driven push messaging bus.  For that reason, email
is good.  Having to poll places for data doesn't scale, having
something come in over the socket and send a kqueue event telling the
listening process to check out the data that's blocking on the socket
and waiting to be processed is good.  Having multiple lists is like
having multiple ports.

Heh... I never thought I'd equate online communities to listening TCP
servers on an IP address.  ::shrug::

> There are any number of ways to automatically fetch data.  Using a
> mail client to pull text from a mail server is just one of them.  If
> people are already making entries to the RAA, why wouldn't an RSS
> feed from RAA be sufficient?  If people *aren't* maintaining entries
> in RAA, should they?  Because if they should, but don't, then why
> would they bother using an announcement list?

I have nothing against the RAA other than that the information format
makes it hard to see what's changed in the course of a week.  That's
what the announcement list will do.

> The sense I'm getting, though, is that one will be able to get some
> information about new libraries from the announce@ list, and some
> (possibly overlapping) information from RAA, but neither one will be
> authoritative.  Instead of having one incomplete source (i.e. RAA),
> we'll have two.  If we want to avoid trolling around for
> information, then having a single authoritative source might work
> better.  If it feeds secondary means of distribution, fine.

It will be a combination of announcements that were sent by the author
(hopefully), or an automated diff (of sorts) of what's changed in the
last week from the RAA.  -sc

-- 
Sean Chittenden