From: Conrad Schneiker <schneik@...>
Date: 2001-01-10T10:53:32+09:00
Subject: [ruby-talk:8971] Re: GULP (working title :-) related thoughts

Dave Thomas wrote:

# David Alan Black <dblack@candle.superlink.net> writes:
# 
# > I've been very interested particularly in the development of an
# > on-line cookbook, as well as the various archive and library ideas
# > that have been circulating recently.  And yet, when I've tried to
# > really sit down and sketch out how things might work, I've hit a kind
# > of impasse.  My latest analysis of this is that what's behind this
# > impasse is, of all things, the very power of Ruby.

Well, I think what you (and the soon-to-be burgeoning world-wide 
Ruby-related job market) eventually/ultimately want is the emergence of a 
hopefully increasingly common web of "best practices idioms" that serve as 
useful reference points and points of departures in the midst of all these 
variances. There is something of a natural development impasse here 
because you can't start out with this, since this would be one of the very 
desirable *products* of such a system as it developed over time. 

# For the cookbook site, as opposed to Gulp, it's
# perfectly acceptable to have many version of things, partial
# snippets, and the like. In fact, think how valuable it would be to
# have all this discussion about Hash defaults archived, along with
# all the code snippets. People could search for 'hash default' and
# get not just working code, but an insight into the process that lead
# to it.
# 
# Don't let the fact that many of the snippets will be trivially short
# be a deterrent. They're educational, and genuinely useful. In fact,
# sometimes the very power of Ruby can obscure a simple solution:
# having all these snippets online and indexed will be invaluable.
# 
# Gulp is a different story. There we're aiming for accuracy,
# completeness and authority. But we can cross that bridge when we've
# worked out exactly what we want.

Gulp doesn't need to be a different story. Think of Gulp as a pyramid. 
There's nothing that precludes a library from having an "industrial 
strength enduring classics" section with rigorous standards and a 
repository section open to all contributors, plus one or more useful 
intermediate levels/types of material and management (including link 
archives). An on-line Ruby cookbook might be an intermediate level in this 
scheme of things. In such terms, the lowest level would be an on-line Ruby 
scrapbook.  You would probably also want a couple of levels under the top 
level for widely used, up and coming packages, that are candidates for 
turning into top-level packages.

Conrad Schneiker
(This note is unofficial and subject to improvement without notice.)