From: Beman Dawes (bdawes_at_[hidden])
Date: 2001-11-05 16:43:29
At 04:28 PM 11/5/2001, Howard Hinnant wrote:
>On Monday, November 5, 2001, at 04:06 PM, Kevin Lynch wrote:
>> So, unless I'm misunderstanding something, I really don't see that it
>> useful to make any changes to the library or language that would loosen
>> the type system in this way to solve what I see as a nonexistent issue.
>> But again, I could easily be missing something, and might well be
>> convinced otherwise.
>You're not misunderstanding. And coming from an engineering background,
>there's a lot of me that agrees with you. However a significant number
>of my customers disagree with us (a few of them pretty strongly). And I
>usually find myself representing them more than my personal opinions on
>cases (like this one) where I see no real harm. This is essentially a C
>compatibility issue. And is targeted straight at the math functions.
>I'm not proposing a general language change.
>Nevertheless, please don't feel obliged to add this issue to your
>newsgroup post if you're not comfortable with it. Since you asked, I
>answered. It need not go beyond that.
Bjarne has been making a strong case for not just documenting rationale for
"paths taken", but also for "paths not taken". Along those lines Kevin
might want to consider keeping Howard's query and his response as part of
the history of issues he forwards to the committee.
OTOH, I can understand why Kevin would not want to muddy the water by
including the exchange of views in his newsgroup posting.
And by the way, thanks for helping!
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk