Home Original page

PEP0238 lament

Craig Markwardt craigmnet at cow.physics.wisc.edu
Mon Jul 23 01:53:34 EDT 2001 
"Tim Peters" <tim.one at home.com> writes:
> [John W. Baxter]
> > I see a fairly strong argument there for starting out with a language
> > in which int / int --> float (or rational) and int // int --> int (or
> > int div int --> int).
> 
> I don't consider a vote an argument.  At best, since it was a visible show
> of hands, it let everyone know who wasn't worthy to live due to voting the
> wrong way <wink>.
> ...

I'll decloak here for a moment.  I'm an IDL user, as in Interactive
Data Language by RSI, but I mostly admire Python from afar, and read
the newsgroup.  I can relate the community experience of the upgrade
between IDL version 4 and version 5.

In the new version, IDL gained an *optional* language change.  IDL 5
allows users to subscript arrays with either the square brackets [] or
the round parentheses ().  IDL 4 only allowed the round parentheses.

As a result of this, people began coding both styles.  However, the
new style was not backwards compatible, so there was some
fragmentation.  Some major standard libraries were, and are still,
maintained in two forms.  

Overall, from introduction to acceptance, the change took about 2-3
years.  This change was thankfully optional.  I am very grateful that
I can still pull up useful IDL archive code from ten or more years
ago, written by someone long gone, and usually have it run
successfully.

Thus, I am dismayed to hear about a Python language change that will
*not* be optional, and which will break the function of existing code.
Doing so will eventually place people in the unenviable quandry of
either upgrading their Python interpreter, or tossing out a tried and
tested library module.

Other language enhancements, like nested scopes, add features at the
expense of resolving language ambiguities.  The proposed integer
division change is an outright language change, and impacts a core
functionality: arithmetic.  Quite frankly, I am baffled.

Craig

-- 
--------------------------------------------------------------------------
Craig B. Markwardt, Ph.D.         EMAIL:    craigmnet at cow.physics.wisc.edu
Astrophysics, IDL, whatever else
--------------------------------------------------------------------------
More information about the Python-list mailing list