PEP 308: Pep Update

[John La Rooy]

On Thu, 13 Feb 2003 00:00:38 -0800
David Eppstein <eppstein at ics.uci.edu> wrote:

> The more this discussion drags on, the less I like any of the proposals.
> The current "and or" trick is looking less good to me, too, but I'm 
> starting to think it's better to place more effort on programmer 
> education about why it's bad than on ways to replace it with something 
> less broken.
> 

Lets see. The "and or" trick saves me about 4(#1) lines of code at the expense
of hiding it's intention from a casual reader.

But at least "and" and "or" have other uses.

It seems silly to me(#2) to add a whole syntactical construct (and maybe a keyword)
to the language just to save 4 lines every now and then.

The cond(C,x,y) function some people have suggested makes sense if you're doing a
bunch of these, so long as you don't need short circuiting.

If you need short circuiting - just use 4 lines. Chances are you're going to need
extra comments to explain what you're doing and now you probably have extra space
at the ends of the lines.

    if C:
        temporary_variable=f(x) 
    else:
        temporary_variable=g(y)
    do_something_with(temporary_variable)

If you need a bunch of short circuiting ternary operators meaning you'll save 
lots of lines. Use 

    cond(C, lambda:f(x) , lambda:g(y) )() 

and have a comment explaining what it does. 



Could we have cond() as a standard function with True and False? Then we could
take the "and or" trick out of the FAQ and tell people "use cond()" instead.
If they ask for a short circuiting form, tell a white lie :o) 

John


#1 Has anyone got an example where more than 4 lines would be saved by a 
   ternary operator?

#2 This is my opinion. You are allowed one too.