Re: Proposal: allow @, $, !, and ? in Lua identifiers

[Andrew Starks <andrew.starks@...>]

On Fri, Apr 25, 2014 at 8:59 AM, Coroutines <coroutines@gmail.com> wrote:
> On Fri, Apr 25, 2014 at 6:07 AM, Andrew Starks <andrew.starks@trms.com> wrote:
>
>> With that data in hand, evaluating this proposal will be enjoyable!
>
> What data would you accept?  I think this is a much different proposal

A narrative of your experience with the changes. How did it go?

> from something people can sample as a module.  Asking others to
> recompile Lua just to try out my changes to identifier naming seems
> unreasonable to me -- most of my friends just use what's packaged by

Why is that? If it is unreasonable to ask you to implement your own
suggestion, then who should do it for you?

> their Linux distro of choice.  I'm not sure many would have the nerve
> to patch Lua ~

I may not understand your meaning correctly. If your friends download
and use Lua, they probably lack an informed opinion about how it
should be changed. After a time, the differences between it and the
languages that they already know, will become apparent. They will
enjoy some differences and hate others, in part because some languages
have real advantages, but *also* because they are understanding Lua
from another language's trade-offs.

Now, it is true that if they happen to be friends that also have the
benefit of decades of experience in language design, then they might
have strong and informed opinions about how the language might be
improved. If they lack that experience, they can have a relatively
uninformed opinions, but can still be delusional about their talent
for spotting gaps.

The only way to know if your friends are delusional or not, is to have
them crack open the source code, hack away and then see if they're
right[1].

>
> I think it's a fairly straightforward proposal, I was mostly looking
> for immediate reasons I might have overlooked for "why not".

Then it's not a proposal. If you're looking for "why not", then you're
asking a question.

> Originally I wanted all punctuation in identifiers (after the first
> character) -- but then I remembered that an operator could follow too
> closely and confuse this.

This is an obvious one. Actually going through the process of
implementing this feature would reveal many not-so-obvious ones.

I might be getting at something here: ideas are precious, when they
are refined by labor. When they aren't, they are not *yet* worth
anything. Because typing is cheap and email is free, putting stuff out
there is too easy to do[2].

-Andrew

[1] For me, there are a couple of things I'd like to try: some kind of
ability to adorn structured objects with type information being the
one that pops to my mind. I've done that already and I like it. Once I
go all of the way with the idea, I'll see if it's useful or not. So
far, given the handful of experiments that I've tried, only two have
survived:

"Hello, %s!"  % {"world"}
--and
local my_table = type.new({}, "foo_type")

print(type.tostring(my_table), type(my_table, "foo_type"))
--> "foo_type", true

The rest of my "holy crap wouldn't it be awesome if..." ideas are in
the garbage dump. Be glad nobody listened to me, either. :)

[2]Perhaps requesting that people try their proposal out, or at least
to implement 1 in 5 of the proposals that they post, would set a
*useful* barrier to the firehose of requests that have been put up on
the list in the last month.