Historical info and a concern...
There used to be a Scheme Cookbook wiki, which was an effort to build
something like a Perl cookbook of the time.
It was a good idea in principle, but two of the problems were:
(1) Most cookbook entries should have been "use library package X, the
introductory documentation of which gives an example that is an exact
solution to this cookbook item title, and we'll copy&paste that example
into the cookbook here, so you feel like the cookbook is giving you
something that you can't get elsewhere even though you can". (This was
a little before PLaneT, but people already distributed non-PLaneT ".plt"
files and individual source code files.) I think that cookbooks are
mostly for code patterns that cannot or should not be made reusable
components, which is a lot more common with a sketchy language like Perl
(e.g., dirty regex tricks that are not easily made components) than with
Racket. Also, general linguistic idioms (e.g., general use for "cond"
"=>"), while not suitable for reusable components, arguably belongs in
documentation of core language instruction, not in a cookbook. That
doesn't leave much for a cookbook, except to be redundant or inappropriate.
(2) IIRC, at one point, there were some low-quality contributions to the
cookbook, due to things like someone making a school project of everyone
having to add something to the cookbook. The wiki software was annoying
to use, plus correcting can be socially awkward, and trying to correct
bad advice was harder than giving better advice in the first place.
A separate consideration to keep in mind when creating any new centers
of Racket information: we already have Racket information in many
places, and it's already getting hard to find because of this
(especially annoying is all the sites mirroring copies of Racket email
lists, so Googling becomes harder, because of duplicates). Racket
documentation set, PLaneT package documentation, documentation not in
the normal Racket documentation set like draft htdp2e, Racket email
lists, info on racket-lang.org that's not in the documentation, official
Racket blogs, Git files that people have started for purposes like lists
of suggested contributions, misc. blogs (including mine), IRC logs,
Hacker News, StackOverflow, 2-3 separate categories Reddit, a wealth of
old stuff on Usenet, potential discussions on social media sites with
Racket groups like LinkedIn, academic papers...
You already need good Web search to manage this scattered info (see
aforementioned problem with all these sites archiving our list), and the
problem simply grows every time someone elects to put Racket information
in yet another place. IMHO, Racket is not yet used widely enough to
explain all this information sprawl as a natural consequence; I think
some of the sprawl might be fragmenting things unnecessarily.
Sometimes this fragmentation might be because the new place for
information is meeting a need not yet met. Sometimes by a desire to go
to the proprietary venues where pockets of users or potential users
might be, rather than welcoming them to come to us. Sometimes to try to
accommodate everyone's preferences/practices, which might indeed be
fringe, and we can say, "Hey, look, we mirror this same information
through all these different channels, so just pick your favorite (just
don't try a Google search, since most of these duplicates go back to the
Web at some point)." Sometimes perhaps not thinking through whether
putting information in such-and-such particular new place makes it very
accessible to all the people to whom we want it accessible. I assume
that there are other categories. There may be good reasons for all of
these, although I assume some are more defensible than others, and that
we don't know all the pros and cons.
Aside on one factor in fragmentation... Keep in mind the context of
proprietary sites, many of which essentially are in the business of
owning slices of community interaction. They would prefer 100%
ownership, but if they can't have that, there's fragmentation when they
grab and hold onto what they can. This ownership is demonstrably valued
at billions or trillions of dollars, and so attempts at it are going to
happen. We have to remember when we're playing the game with someone
who wants to own a part of us, and accept the costs of fragmentation
when we do.
(I'll catch up on reading people's further comments in the next day or
so. I had to type this out really fast, since I'm just "on break" from
a work crunch.)
Neil V.
--
http://www.neilvandyke.org/
____________________
Racket Users list:
http://lists.racket-lang.org/users
