Deposing Dictators

Tim Peters tim.one at home.com
Wed Aug 15 10:47:23 CEST 2001


[Arthur Siegel]
> ...
> Guido, after misdirecting the discussion

Prejudicial choice of words, don't you think?

> by implying he was going to pass on the PEP due to the code breakage
> issue,

That was his view at the time; it changed later.

> announces he is sending out a patch previewing the new division
> semantics to the desperately needy VPython folks

"desparately needy" is more counterproductive excess.  Since they and Alice
had the longest sustained visible interest in this issue, sure, of course he
made the patch available to them.  It's equally available to you.

> - PEP accepted.

Not yet at that point.

> At the time, about all the PEP said was that the division semantics
> of Python was a major stumbling block for the newbie.

Guido had not written any of the PEP by that time; it was more like Moshe's
cryptic notes to himself; Moshe doesn't speak for Guido, and it was Moshe's
PEP then.  What changed in the meantime was the full realization that (as
the PEP now makes clear) unification of the numeric model would also be
blocked without changing int division, and Guido took over the PEP then,
rewriting it from scratch.  There are several reasons to change division,
and it's the confluence of independent pressures toward the same end that
drove it to critical mass.

> Guido responded to one of my own early posts by throwing the VPython
> experience up as a proof of the significance of the issue.
> Bruce Sherwood's report to him on the VPython's students experience
> was quoted at length.

What of it?  Nobody is claiming that the VPython folks don't want this
change -- and reports based on experience instead of pure rhetoric have a
certain charm due to their extreme rarity <0.9 wink>.

> The obvious line here -
>
> You can rewrite the PEP, but not the history.

I'm not trying to.

> The only significant information that has been revealed in this
> exchange is again no particular surprise to me.  You linked a few
> posts ago to a long Bruce Sherwood post to Python-Dev about the
> significance of the div issue for his use of Python in his physics
> class.

It was actually to Idle-Dev.  Now you can get worked up about how Bruce was
trying to hide it from the Python developers <wink>.

> The post was *before* he had used it in the classroom.
> He had pre-determined it as a problem - not experienced it
> as a problem.
> ...

Sorry, I still don't see what it has to do with PEP 238.  Even assuming it
were true, which I don't -- while Bruce may not have used Python in a
classroom before then (the referenced post doesn't really say), he had many
years of prior classroom experience, and was one of the designers of the cT
programming language, and used cT in the classroom.  The idea that his
expressed convictions were baseless is thus one I don't buy.

BTW, cT has three distinct division operators:  "/" (always floating),
"$div$" (truncating to int) and "$divr$" (rounding to nearest int).  You can
argue with him about whether cT's decision to supply those damaged his
students, but it's not relevant to PEP 238 regardless.

outta-this-ly y'rs  - tim





More information about the Python-list mailing list