A critic of Guido's blog on Python's lambda

Thomas F. Burdick tfb at conquest.OCF.Berkeley.EDU
Sat May 6 21:02:26 CEST 2006


[ I pruned the cross-posting down to a reasonable level ]

Ken Tilton <kentilton at gmail.com> writes:

> Thomas F. Burdick wrote:
>
> > This is second-hand, as I don't actually follow Python closely, but
> > from what I've heard, they now have reasonable scoping rules (or maybe
> > they're about to, I'm not sure).  And you can use def as a
> > Scheme-style inner define, so it's essentially a LABELS that gets the
> > indentation wrong.
> 
> Cool. And I know how much you like labels/flet. :)

Well, I love LABELS but I hate inner defines with an equal passion --
so for me it's a wash :-)

As much as I like nice low-level, close-to-the-machine mechanisms as
labels and lambda, sometimes you just want the high-level
expressiveness of tagbody/go, which Python doesn't have ... which in
my opinion is quite a crime to readability and the ability to
transcribe Knuth algorithms, which any engineer should find offensive
to their sensibilities.

> >  This means they have proper closures, just not
> > anonymous ones.  And an egregiously misnamed lambda that should be
> > fixed or thrown out.
> > If Python gets proper macros it won't matter one bit that they only
> > have named closures, since you can macro that away in a blink of an
> > eye.
> 
> Ah, well, there we go again. Without sexpr notation, the lexer/parser
> again will be "hard", and "hardly worth it": we get even more sour
> grapes, this time about macros not being such a big deal.
> 
> One of the hardest things for a technologist to do is admit that a
> neat idea has to be abandoned. Initial success creates a giddy
> over-commitment to the design choice. After then all difficulties get
> brushed aside or kludged.

Y'never know, they could always Greenspun their way to almost-sexps.
What with the way that selective pressure works, it's gonna be that or
die, so it is a possibility.



More information about the Python-list mailing list