Deposing Dictators

Arthur Siegel ajs at ix.netcom.com
Tue Aug 7 14:55:45 CEST 2001


As Stephen and I seem to have in common some deep resentments
over a small language change, I thought it would be interesting to 
isolate *our* differences.

Stephen writes :

>If there was an undeniably good reason for the change, I would agree
>with this. As it stands, there are equally good - if not better -
>reasons for keeping division as it is.

I personally have no strong feeling about how a div operator should
work or whether a language should be case insensitive.  

But it is clear to me that the issues have been given unreasonable
significance by those with the burden of proof to justify a
change from the status quo.

Why?

It is also  clear to me that the div operator issue argument 
could not, would not, and should not have gotten as heated 
as it did were it not for the backdrop of the case insensitivity
threat.

>At the end of the day, the change is being done because of a religious
>viewpoint held by a vocal part of the group that has a long history
>of rioting at the palace gates. The decision isn't being made by
>logic or by popularity - it's being made because the DFL has been
>hearing the protestors shouting for too long while the happy and
>indifferent users have just got on with using Python. He is a new-born
>division-must-be-done-this-way-ist.

I disagree with Stephen here. This was a top down issue
always - not a reaction to the clamoring of a vocal minority.
Guido himself seems to have a religious perspective on what
CP4E is about - a vision thing. Strongly influenced by Alice, the
CMU Virtual Reality project. Suffice it to say that IMO it is a 
quirky vision.

Have ranted at length about it on EDU-SIG. It is the Alice project
that confirmed the significance of case sensitivity to Guido. I and 
others tried to convince Guido that those findings were flawed to 
such an extent where they could be safely ignored. Tim Peters, 
if I am not mistaken, publicly weighed in on it- "flawed". 
Deaf ears.

Then VPython comes along - also out of Carnegie Mellon.  Guido 
states that VPython "failed to confirm" Alice's findings about the 
significance of case sensitivity.  I note Guido's wording. In better 
words, "contradicted" those findings. No surprise.  

Stating it as diplomatically as I can, the argument 
that anything of significance could be generalized from
Alice's findings was, is, and has been flawed. 

Rather than going back to the drawing board and questioning 
himself about where he is looking for evidence, and how 
he is evaluating it - Guido now takes VPython's findings 
about the "/" operator (and its "confirmation" of Alice's 
"findings") as the new testament. 

Or at least seems to expect us to be satisfied with it.

I happen to remained convinced that any conclusions drawn
from the VPython reports are simplistically linear as to
its implications for the inexperienced programmer working
in the numerical realm within the overall framework of Python. 

Realize that VPython is Numeric through and through.  The
distribution includes it. All its demos draw from it.

As per the recent Zero Division Error "bug report" on the 
list.

import Numeric
a=array([1,2,3]
a[0]=.4
1/a[0] = Zero Division Error

With 3/4 = 0, the clearest signpost to Python's numeric
behavior, gone - Numeric behavior does in fact look more
and more bizarre and mysterious and unreasonable.

The fact is that next semesters bugs produced against VPython
will be deeper, harder to fathom and uncover.  Students
will have a better basis to complain of having been trapped
by design flaws of the tool they have been assigned to use
then they had last semester IMO.

ART








More information about the Python-list mailing list