[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: Lexical byte-compilation warnings cleanup

From: Drew Adams
Subject: RE: Lexical byte-compilation warnings cleanup
Date: Tue, 20 Aug 2013 19:57:49 -0700 (PDT)

>  >>>> Indeed, it does not refer to the dynamically bound variable.
>  >>> Why is that? Will this be fixed, or is this the intended design?
>  >> Intended design.
>  > And the intention is?  The design is?  The reason is?
> Give it up, Drew.  RMS has never been one to prefer "a foolish
> consistency"

No one has argued for foolish consistency.  Or even for consistency
in this context.

> (ie, adherence to abstract principles)

Or for adherence to any abstract principles. 

> where consistency collides with his intuition or hacking convenience,
> and that same reliance on intuition appeals to current maintainers
> as well:

Nor has anyone argued against using intuition.  That can sometimes
be helpful.

But FWIW I don't think that RMS was particularly one to shy from
presenting his design choices explicitly and giving reasons for them.

Anyway, that was then and this is now - I don't see much point in
comparing with RMS here, either wrt the design or the process.  It's
not clear that RMS would have moved so quickly and cleanly toward
integrating lexical scoping - and I'm glad that has been done.

(No, sorry for the passive voice - I'm glad that Stefan did that.)

But I would like to understand more (more than zero, at least) about
the design intentions in this regard, and the reasons.

Since there are 35 years of big brother Common Lisp's experience
mixing lexical and dynamic binding in the same language, not to
mention years of high-level discussion by Lisp implementors of all
stripes to come up with that design in the first place, I would think
that doing something quite different would prompt at least some
discussion of alternatives, pros & cons, etc.

That was admittedly a long time ago, and no doubt knowledge and
technique have progressed since then.  All the more reason why I
would like to know something about what is behind this new approach.

Aren't you curious at all why this difference wrt function parameters?
That possibility never even occurred to me.  I'm sure there are some
reasons for the choices made, but they don't seem to be forthcoming.

>  > > As language designers, we just take idea from here and there.
> I think that is the last word on Elisp design principles.  Abstractly
> I don't like it, myself, but surely you don't deny that it has proven
> to be a workable strategy.

Again, one can take ideas from here and there - no problem with that.
But one takes this idea instead of that one for a reason.  And that's
what I'm asking about.

Is it a good thing or a bad thing for more people to have an idea
what the design is about and what motivates it?

> Nevertheless there is hope for those who prefer care-full design:
> Guilemacs!

For the record, Guilemacs is not my hope.  (Not that I'm sure what
you mean by that term.)

So far, the addition of lexical scoping to Emacs has not been far
from the way it is mixed in with dynamic binding in Common Lisp,
AFAICT.  Naively, I would hope for Emacs Lisp to become more similar
in this regard, not more different.

But I'm no expert on any of this.  Which is why I asked the question.
And that's all I've done: ask why.

I ask it again: why?  The answer so far is to enable efficient
interpretation.  Sounds like that might be a promising reason, but
so far that slogan means nothing to me.

Does it mean something particular to you?  Don't you want to know
more about what is intended, and why, wrt this design which is
presumably so important to Emacs and Emacs-Lisp programmers?

reply via email to

[Prev in Thread] Current Thread [Next in Thread]