PEP 308: Pep Update

[Peter Hansen]

Raymond Hettinger wrote:
> 
> Amended the PEP to reflect some convergence on the newsgroup:
> 
> The leading options on the table are:
> 
> *     (if <cond>: <expr1> else: >expr2>)
> *     <cond> then <expr1> else <expr2>
> *     <cond> ? <expr1> : <expr2>
> *     no change

My main problem with some of the proposals, including the one
above with "then", is that they do not seem to me to suggest that
there is a condition and a selection.  In my mind, to do that
requires a question to be asked, as in "if" or the C-like "?".

The word "then" by itself has merely a connotation of sequential
time, so that "A then B" implies that both A and B will happen
in some way, with B following A in time.

Without it's partner "if", "then" fails to convey the question
that is being asked.

I also have a mild concern about the first contender listed above,
because in my mind the colons are for introducing blocks, while
in the above they are "overloaded" merely to help the syntax be
unambiguous (I assume), or for some other reason not involving
blocks.  

I think I could live with that, although at this point I'm starting
to think that only Guido's original ("backwards") suggestion is
elegant, though still undesirable to me, so I'm now favouring 
the final option above again... sigh.

-Peter