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

Alex Martelli aleaxit at yahoo.com
Sat May 6 09:13:49 CEST 2006


Ken Tilton <kentilton at gmail.com> wrote:
   ...
> > Absolutely.  That's why firms who are interested in building *seriously*
> > large scale systems, like my employer (and supplier of your free mail
   ...
> > Obviously will not scale.  Never.
> > 
> > Well... hardly ever!
> 
> You are talking about being incredibly popular. I was talking about 

Who, me?  I'm talking about the deliberate, eyes-wide-open choice by
*ONE* firm -- one which happens to more or less *redefine* what "large
scale" computation *means*, along many axes.  That's got nothing to do
with Python being "incredibly popular": it has everything to do with
scalability -- the choice was made in the late '90s (and, incidentally,
by people quite familiar with lisp... no less than the reddit.com guys,
you know, the ones who recently chose to rewrite their side from Lisp to
Python...?), based on scalability issues, definitely not "popularity"
(Python in the late '90s was a very obscure, little-known language).

> kenny (who is old enough to have seen many a language come and go)

See your "many a language" and raise you one penny -- besides sundry
Basic dialects, machine languages, and microcode[s], I started out with
Fortran IV and APL, and I have professionally programmed in Pascal (many
dialects), Rexx, Forth, PL/I, Cobol, Lisp before there was a "Common"
one, Prolog, Scheme, Icon, Tcl, Awk, EDL, and several proprietary 3rd
and 4th generation languages -- as well of course as C and its
descendants such as C++ and Java, and Perl. Many other languages I've
studied and played with, I've never programmed _professionally_ (i.e.,
been paid for programs in those languages), but I've written enough
"toy" programs to get some feeling for (Ruby, SML, O'CAML, Haskell,
Snobol, FP/1, Applescript, C#, Javascript, Erlang, Mozart, ...).

Out of all languages I know, I've deliberately chosen to specialize in
Python, *because it scales better* (yes, functional programming is
_conceptually_ perfect, but one can never find sufficiently large teams
of people with the right highly-abstract mathematical mindset and at the
same time with sufficiently down-to-earth pragmaticity -- so, for _real
world_ uses, Python scales better). When I was unable to convince top
management, at the firm at which I was the top programmer, that the firm
should move to Python (beyond the pilot projects which I led and gave
such stellar results), I quit, and for years I made a great living as a
freelance consultant (mostly in Python -- once in a while, a touch of
Pyrex, C or C++ as a vigorish;-).

That's how come I ended up working at the firm supplying your free mail
(as Uber Tech Lead) -- they reached across an ocean to lure me to move
from my native Italy to California, and my proven excellence in Python
was their prime motive. The terms of their offer were just too
incredible to pass by... so, I rapidly got my O1 visa ("alien of
exceptional skills"), and here I am, happily ubertechleading... and
enjoying Python and its incredibly good scalability every single day!


Alex



More information about the Python-list mailing list