Deposing Dictators

[Dirck Blaskey]

James Logajan <JamesL at Lugoj.Com> wrote in message news:<3B61B01F.6C6B2B5D at Lugoj.Com>...
> ...
> I suspect Mr. van Rossum is probably not in the happiest of moods, what with
> his suggested changes being challenged from many people for various reasons.
> This influence would color anyone's reading.

I would wager it's not the challenge itself that's wearing.

Any rational discussion of a technical disagreement is appropriate -
but the response to this change has been completely overboard.

I haven't followed it all, because 
1) A lot of it is just plain rude
2) There's simply way too much of it
3) There doesn't appear to be much in the way of meaningful argument

Guido doesn't deserve a lot of the crap thrown his way over this,
and it wouldn't surprise me if he told us all 'Piss Off, I'm going fishing'.

It's much harder being benevolent dictator than despotic ruler.
(and: "Since when is the majority ever right, anyway?")

So far I've managed to avoid contributing, thinking that far, far too 
much has been said on this topic - I'm tempted to invoke Godwin's Law.

But now I'm gonna open-my-mouth-too-and-insert-both-feet,
basically in the 'cooler-heads-will-eventually-prevail' department.

In the end, it's just code.  No reason to shout and get red in the face.

> By the way, the division issue is several years old. A lot of bits have
> flowed on this issue. See for example:

> http://groups.google.com/groups?hl=en&safe=off&th=1d9d3c24ba2f933d,40&start=0

> Three years later the code base and programmer base are no doubt
> considerably larger, so the inertia against change has probably grown too
> large to deal with.

When I first heard about this back then, I had a 
knee-jerk-dogmatic-negative-reaction to it.  Now, I believe
I understand the motivation for the change better, and I also
understand where the reaction comes from.  I've also had direct
experience with a co-worker being burned by integer division.
He didn't know what the problem was until I pointed it out to him.

Guido's motivation is pretty simple:  sand off the rough spots.
Prevent a problem from occuring instead of hand-waving it away after.

("Just cut away all the stone that doesn't look like the statue.")

The arguments against come down to two groups:

1) Inertia, two flavors:
        
        a) Too Much Code will be Impacted
                (and The Children will be Corrupted!)
        
        b) This Displays Wanton Disregard for Stability that will 
                Desperately Frighten the Management 
                (obviously from the non-Unix and non-Windows users ;)
                ("How long do we still have to support SunOS 4.1.x?")

Inertia should always be taken into account, but it should never
be the single deciding factor - or we'd all be working legacy COBOL.

Guido has taken great pains to address this issue, because, technically,
it really is the only issue - everything else is just reactionary.
While I wouldn't argue against leaving the old opcode available indefinitely,
Guido is more meticulous than that.

2) The unspoken 'do it the C-way' prejudice.

I think this is where most of the reactionary responses come from.
C is deeply ingrained in the programming culture, because in a lot
of ways it was, and in some ways still is, The Programming Language.
Anything you do, day in and day out, year after year, no matter how 
odd or perverse or inconsequential, becomes 'the natural order of things'.
The more deeply ingrained, the harder it is to actually see.

Sometimes you have to get down to the bits, somewhere close to the hardware.
I don't see this going away in Python - it's just trying to move
the hardware issues out of the way when they're not relevant.
Python will always integrate well with C, because that's part of its nature.

...

I just wanted to add a thank-you to Guido, for a terrific language.
It really is a joy to write Python code.
 
d