subexpressions

[Paul Boddie]

Re: subexpressions

Steve Howell wrote:

>
The compiler doesn't know the types up front, but if
you wanted to do this kind of optimization (and you
believed that 95% of x*x cases would benefit from it,
and you're willing to sacrifice performance for the 5%
of folks that overload multiply), then the compiler
could generate bytecode that set the stage for later
conditional caching of the first execution of x*x.

True.

You'd then need the execution of the bytecodes at
runtime (ceval.c or something called by it) to work in
such a way that they only cache the result when side
effects are not an issue. At runtime you can reliably
detect whether something is still a virgin builtin,
correct?

I've no idea, but I imagine that psyco knows whether or not it has a
proper built-in number object when it generates specialised code for
similar cases.

To my disclaimer, you would only undertake such an
optimization if multiplication were really, really
expensive (which I don't think is even true for floats
today), and even then you'd proceed cautiously.

Indeed. Some believe that for "full Python" you can only introduce
such measures at run-time, although extensive enough analysis of the
code could perhaps suggest suitable specialisations in advance, as
presumably demonstrated by Shed Skin.

Paul

[Steven D'Aprano]

Re: subexpressions

On Fri, 01 Jun 2007 07:09:50 -0400, Steve Holden wrote:

>>>>The real answer is of course: Use a function.
>>>But what about something like
>>>>
>>>lambda x: sin(y)+cos(y) where y=x*x
>>>>
>>>?
>>>May be this could be a PEP? If there is no straight way to do this.
>>def f(x):
>> y = x*x
>> return sin(y) + cos(y)
>>>
>>What is not straightforward about that?

>>
>This code is needed once in a map, so I don't want 3+ extra lines.
>Solution seemed so simple...
>I always considered python as languague, where simple things do not require
>extensive coding.
>Moreover, this construction is common thing in functional programming.
>>
>>

Stop thinking of three lines as "extensive coding" and your problem
disappears immediately.

The F-bot once suggested adding a clause to the Zen of Python about
"writing two lines of code is not a sin" or "cramming two lines of code
into one is not a virtue" (my paraphrases).


Check the two alternatives:

def f(x):
y = x*x
return sin(y) + cos(y)

44 key presses, including tabs and newlines and a blank line after the
function, but excluding counting the shift key separately.

lambda x: (lambda y: sin(y) + cos(y))(x*x)

42 key presses.

Apart from the extremely minor issue of "namespace pollution", I think
that speaks for itself.



--
Steven.

[Cousin Stanley]

Re: subexpressions

....
After years of discussion, Guido has decided
to leave lambda alone for 3.0.
>
It will not be neither expanded, nor removed, nor renamed.

But it still will be as ugh, ugh, ugh-lee
as a mule walking backwards ..... ;-)


--
Stanley C. Kitching
Human Being
Phoenix, Arizona


----== Posted via Newsfeeds.Com - Unlimited-Unrestricted-Secure Usenet News==----
http://www.newsfeeds.com The #1 Newsgroup Service in the World! 120,000+ Newsgroups
----= East and West-Coast Server Farms - Total Privacy via Encryption =----

[Terry Reedy]

Re: subexpressions


"Cousin Stanley" <cousinstanley@ hotmail.comwrot e in message
news:1180756858 _54747@sp12lax. superfeed.net.. .
|
| ....
| After years of discussion, Guido has decided
| to leave lambda alone for 3.0.
| >
| It will not be neither expanded, nor removed, nor renamed.
|
| But it still will be as ugh, ugh, ugh-lee
| as a mule walking backwards ..... ;-)

Then pretend it was eliminated, as Guido once thought to do, and do not use
it. And look away when others do ;-)

tjr

[Stef Mientki]

Re: subexpressions

>

>
Check the two alternatives:
>
def f(x):
y = x*x
return sin(y) + cos(y)
>
44 key presses, including tabs and newlines and a blank line after the
function, but excluding counting the shift key separately.
>
lambda x: (lambda y: sin(y) + cos(y))(x*x)
>
42 key presses.
>
Apart from the extremely minor issue of "namespace pollution", I think
that speaks for itself.

and now I've only 60 lines on my screen,
so what about

def f(x): y = x*x; return sin(y)+cos(y);

cheers,
Stef Mientki

[Alex Martelli]

Re: subexpressions (OT: math)

Stebanoid@gmail .com <Stebanoid@gmai l.comwrote:

On 3 , 22:07, "Steban...@gmai l.com" <Steban...@gmai l.comwrote:

angle is a ratio of two length and
dimensionless.http://en.wikipedia.org/wiki/Angle#U...easure_for_ang
les

only dimensionless values can be a argument of a sine and exponent!
Are you discordant?

>
if you are discordant read more :P :
sine is a dimensionless value.
if we expand sine in taylor series sin(x) = x - (x^3)/6 + (x^5)/120
etc.
you can see that sin can be dimensionless only if x is dimensionless
too.
>
I am a professional physicist and a know about what I talk

Lots of people are confused by the concept of "degrees" -- your Taylor
series, of course, intrinsically assumes x is "in radians" (which of
course IS how angles "truly are"). I blame the Babylonians for that
confusion just as much as for the clunky base-60 that intrudes in our
ordinary time reckoning...!


Alex