Jonathan S. Shapiro wrote:
So far as I can tell, I am making only one of the assumptions that you
have enumerated. I cannot speak for anyone else in the group.

I did not mean to imply that any one individual held all of these assumptions.


[...] Rather than ascribe opinions,
perhaps it would be more appropriate to simply state what *you* believe.

Agreed. I see no reason to further examine the issue of whether these other assumption have been implicit in some of the messages so far posted to this thread. In any case, I hope my message was clear about what I do believe.



* On Pentium architectures, in the absence of proof that we're safe from overflow, the overflow check can only be done by an explicit check, which is a cost that will deter C programmers from switching to BitC.

My take: These same C programmers will be deterred by the bounds check on array indexing...

This is an assumption that I am actually making. The difficulty with your rebuttal is that my "assumption" is amply supported by empirical evidence. Given which, I am reluctant to take a final position on the behavior of integer operators until I understand the difficulty of discharging the "does not overflow" lemma.

So how do you evaluate my modest proposal, as a way to make BitC more palatable to C programmers? Assuming you dismiss it, as I would hope, what difference do you see between the "+" issue and the indexing issue?


--
Text by me above is hereby placed in the public domain

    Cheers,
    --MarkM

_______________________________________________
bitc-dev mailing list
[email protected]
http://www.coyotos.org/mailman/listinfo/bitc-dev

Reply via email to