Seems to me Stepanov bailed out of this discussion when the going got
tough.
Certainly C++ has certain properties that other OO languages don't
have, but nobody seems to have pointed out any that are necessarily
that desirable, let alone necessary to implement his ideas.
Probably the only reason he chose C++ was an ill-conceived belief that
it provided "hand-coded performance". He was probably lulled into this
false premise because C++ is so "close to the metal". That's
understandable I guess, many have made the same error.
Perhaps another possibility is that he actually discovered these
principles while hacking lisp, and then noticed that they could be
applyed to C++, but nobody had yet bothered. A good opportunity to
make one's fame and fortune so to speak. It might have even succeeded
as a universally acknowleged Good Thing, if it wasn't for the
fundamental disaster that C++ is.
So at the end of the day I don't know whether to admire him for
extending the limits of generic programming in a static typed
language, or to ridicule him for pretending he invented some new form
generic programming that can only be implemented in C++.
From: Paul Campbell
Subject: Re: What Stepanov *really* meant [Was: STL efficiency]
Date:
Message-ID: <33981574.ABD322C@lucent.com>
·····@telstra.com.au wrote:
>
> Seems to me Stepanov bailed out of this discussion when the going got
> tough.
>
> Certainly C++ has certain properties that other OO languages don't
> have, but nobody seems to have pointed out any that are necessarily
> that desirable, let alone necessary to implement his ideas.
>
> Probably the only reason he chose C++ was an ill-conceived belief that
> it provided "hand-coded performance". He was probably lulled into this
> false premise because C++ is so "close to the metal". That's
> understandable I guess, many have made the same error.
Prehaps he chose C++ because he wanted his work to be of real benefit
to 10's of 1000's of practicing programmers rather than rotting away in
some University library.
Paul C.
UK.