Re: [ANN] Microlight (was Re: paving cowpaths with libraries)

On Thu, Feb 16, 2012 at 4:43 AM, steve donovan <steve.j.donovan@gmail.com> wrote:

On Thu, Feb 16, 2012 at 11:02 AM, marbux <marbux@gmail.com> wrote:
> flag candidates for improvement of function documentation, in
> particular with beginning coders and Lua newbies in mind.

This is true. Just having some LuaDoc-style function documentation is
not sufficient. There has to be fuller explanation ('narrative
documentation') and justification. I'm working on a little article and
I'll downplay some of the functional tricks that entertain me.

Generally newcomers come with expectations. They may be frustrated
that that they cannot just print out a table in the interactive
prompt. Or they're accustomed to splitting strings and having
variables expanded in them. Lua string patterns are going to be new
and thus they are going to be scary (and even regexp-familiar people
need to adjust their expectations).

The most tricky kind of expectation is that every language must have a
'class' keyword. That's a big stumbling block. Having something that
works as expected allows them to apply previous experience.

Well, we understand that these newcomers may put these training wheels
behind them later.. But a little bit of initial support is not a bad
thing, and most of that support is explanation and examples. A few
little functions could make a difference.

I don't think we're shopping for a standard library here; there's this
new sport called 'code golf'; this particular little exercise is more
like 'library golf' - see how much useful stuff we can pack into a few
functions.

steve d.

I have been quietly following this thread, but feel like I should chime in a bit.

I don't understand why Penlight/stdlib is not the "cowpath"? I use Penlight most so that is what I am familiar with, so that is where my comments will be focused. Penlight has two levels, but ultimately I feel like users are looking for a single library to fill in the gaps for application creation. This is where Penlight shines. It has high level application code like a command-line parser, down to config file reading, all the way to algorithm/design pattern code. Is it the fact that it has many files? I am trying to get a good sense for the reason for all this. Thoughts?

Do people think that if we make a single file library it will magically sneak into the official Lua source release? I think people should focus on full distributions like LuaDist[1], Lua for Windows[2], and LuaRocks[3]. In fact, if you look at the official Lua home page there is a link to Lua for Windows. I know in other threads they have mentioned they need at least a link on the official page to be considered "standard", thoughts?

I am not trying to stir people up, just get a true direction for what people want and if anything current can fill that requirement with a bit of help.

Thanks for engaging me.
--

Regards,
Ryan