At long last...

Carl Banks pavlovevidence at gmail.com
Fri Jun 20 22:07:38 CEST 2008


On Jun 19, 11:26 pm, Dan Bishop <danb... at yahoo.com> wrote:
> On Jun 19, 9:24 pm, Carl Banks <pavlovevide... at gmail.com> wrote:
>
>
>
> > On Jun 19, 10:17 pm, Terry Reedy <tjre... at udel.edu> wrote:
>
> > > Carl Banks wrote:
> > > > Tuples will have an index method in Python 2.6.
>
> > > > I promise I won't indiscriminately use tuples for homogenous data.
> > > > Honest.  Scout's honor.  Cross my heart.
>
> > > Use them as you want.  This change came about because .index was
> > > included in the 3.0 Sequence ABC (abstract base class) and tuple was
> > > included as a sequence, so .... something had to give.  The result was
> > > tuple getting the full suite of immutable sequence methods.  And then
> > > there was no good reason to not backport ;-).
>
> > The last time I needed index on a tuple was in fact for partially non-
> > homogenous data.  I forget why, but I needed to treat arguments after
> > a certain value different from the front arguments.  So I wanted to do
> > something like:
>
> > def something(*args):
> >     firstspecial = args.index(0)
>
> > 'Cept I couldn't.
>
> Why didn't you just use a list inside the tuple?

I don't remember, but knowing myself I was probably implementing some
sort of declarative function that I would call 200 times with hand-
entered data to build some kind of dataset.

Normally I would not design functions like that, but when "ease of
data entry" is the predominant concern I will do all kinds of hacky
stuff.


Carl Banks



More information about the Python-list mailing list