Re: Project lead nominations for standard libraries?

[steve donovan <steve.j.donovan@...>]

On Fri, Dec 31, 2010 at 5:17 AM, KHMan <keinhong@gmail.com> wrote:
> But thinking up an "approved" proposal and implement it and expecting the
> market to swoon is really too smooth-sailing. I don't believe in anointed
> widgets -- it has to earn its place after being put into practice. No such
> thing as a guaranteed result.

I have to agree with this - designing libraries by committee is a
tedious process that does not guarantee anything.  As you say, this
theme reappears every couple of years, and we never get very far.  I
like Cosmin's idea of 'Darwinian Selection' for functions, although
anybody who has written a general-purpose library knows that a library
is more than a collection of independent functions.

Starting from that approximation (a library is a collection of
functions) gets us something like luacode.snippets.org.  The idea
there was to build up a catalog of useful recipes so that there was
less re-invention (splitting a string into a array is a common
operation, but a little tricky to get right).  Now I'm seeing
HyperHacker's comment below on the usefulness of wikis for this
purpose - the above site is an example of using an intelligent wiki
(Sputnik) and trying to capture the information in a _structured_ way
- author, licence, comments, code, etc.

 The Lua users wiki is a fantastic resource and was an important part
of my Lua education, but it does not self-organize and the fact that
it is in a good shape is due to the silent labour of volunteers like
David Manura and John Belmonte. (A few tweaks would be useful, like
ToCs and intra-document links, but that's a topic for another
discussion.)

steve d.