List Comprehensions Enhancement

[William Tanksley]

On Mon, 13 Sep 1999 22:37:44 -0400, Tim Peters wrote:

>[William Tanksley, comes to the right conclusion]
>> I actually liked the verbosity a lot.  I thought it was extremely cool;
>> I'd like to see this in Python, although I suspect it won't make it (it's
>> a little too much like a functional language).

>This was debated at extraordinary length and in extraordinary detail last
>year (or last decade, or some time span long enough ago that I can't
>remember quite when -- maybe last week).  Guido was kindly disposed toward
>it, and for the same reason Haskell fans can never go back to Scheme <wink>:
>it can replace almost all common uses of map and filter with much nicer
>syntax; and, in Python's case, can be implemented much more efficiently than
>map/filter are implemented today, via generating inline nested loops.

I can understand the possibility of greater efficiency (it's conceivable
that we'd be able to use C to speed it up in some cases); it's also very
cool that it's available _now_.

>So, now that Greg has stopped harassing Guido in his sig <wink>, it stands a
>better chance than you may think.

I didn't notice Greg doing that.  How DOES one harass Guido, anyhow?  I
can't think of anything to say about him that isn't a compliment -- "Guido
is such an absolute dictator!"  "Guido never listens to my suggestions!"
and so on.

>>> Also, I think there should be a clear separator between the
>>> generators/predicates in your syntax too:

>>> print [(i,s) for i in nums, for s in strs]
>>> or similar.

>> I agree -- my initial reaction was to put 'and' as the seperator.  Amazing
>> how great minds think alike :).

>A good way to kill it would be to reopen debate about the syntax <0.9
>wink -- but finding a thoroughly backward-compatible and unambiguous
>spelling isn't easy, and off-the-cuff "gosh, I like the way *this* looks
>better" isn't going to succeed>.

That's true.  It's also true that my 'and' suggestion won't work -- it
causes ambiguity:

[(i,s) for i in pairs if i[0] != 3 and i[1] != 4 and for s in strings]

Comma looks fine; let's see...

[(i,s) for i in pairs if i != (3,4), for s in strings]

I don't see any ambiguity.  OTOH, there's no ambiguity with no
punctuation, either.  I'll be glad to see it with anything except
nonexistance.

>>> Nice thing, though :-)

>> Oh, amazingly.  Thanks, Greg.  On a scale of one to ten I'd give it a tim.

>Oh, it's not *that* good <wink>.

As Quayle once might have said, "I stand by all the misstatements I may,
or may not, have made."

>but-it-is-more-pythonic-than-python-ly y'rs  - tim

-- 
-William "Billy" Tanksley