On 15/09/2010 10.49, James Kanze wrote:
On 2010-09-14 12:13:51 -0400, kwikius said:
[...]
If the intended implication is that big companies somehow subvert the
C++ standards committee so that they can promote their own agendas,
that's simply false. For example, Microsoft really, really, really
wanted resumable exceptions back in the 90s, but didn't get them.
It's probably worth pointing out that within ANSI, each member
has a single vote. And that if your employer is a member, you
cannot be a member as an individual (which prevents large
companies from "stuffing" the ballot box). The result is that
you probably have more impact as an individual member than as an
employee of a large company. (But a lot depends on the
individual. Herb Sutter obviously has a lot of influence. But
because of the books he wrote, not because he works for
Microsoft.)
Well, he is very capable but if the reason why he is "followed"
is that he wrote books...
Not just that he wrote books, but because of the books he wrote.
(Schildt also wrote a lot of books, but his influence, even if
he tried to exercise it, wouldn't be the same.)
Ah, the kind of books, you meant (their quality). Not sure if
your sentence was really ambiguous or my bad English stroke
again.
I wouldn't call it bad English: it's a subtle issue. But
"because of the books he wrote" is not the same thing as
"because he has written (many) books". It's the quality of the
books which makes him respected, not the fact that he's written
many.
Wouldn't Italian have the same distinction: "a causa dei libri
che ha scritto", rather than "perche ha scritto libri". (Not
sure: my Italian isn't that good. Although if I think about it,
the French is somewhat more ambiguous than the English. And the
English isn't totally unambiguous: "because of the quality of
the books he wrote" would be much clearer.)
[...]
Do you have the document number of your proposal(s)?
No. It was several years ago, when I had just started with C++,
and it must have been on one of the newsgroups. At the time, the
replies from people who were in the committee discouraged me and
I didn't make any proposal (note that I put "response" in
quotes, above). But that's bad, because --as I've had a chance
to see later-- it might have gone on anyway.
So someone criticized the proposal in a newsgroup, and you
consider that a rejection by the committee. That's not the way
it works, and the most influential members of the committee
rarely post or comment here.
A similar issue concerned a keyword to denote null pointers. It
seemed that you were a fool just because you were proposing to
add one (hence the 0p tradeoff I mentioned earlier). But now we
have nullptr. Really? Did we say it's a bad thing to add
keywords?
(Not to talk of consistency, spelling out names in
full --such as null_pointer-- etc. But then, don't we have
showmanyc(), epptr() and all that?). So, you see: do as we say,
not as we do.
The problem was more or less officially recognized in the
response to a proposal I made for C++98. The committee's
response at the time was that my proposal didn't go far enough
(it was a halfway measure), that it was too late to consider
a more complete proposal (which was true at the time, which was
why I only made a halfway proposal), and that the issue should
be addressed in a future version.
In other words, the committee had more or less decided that
nullptr (or something similar) would be present in the future.
(Note that it took several revisions to get nullptr working.
It's far from trivial.)
You've not been to many of the meetings, and heard the
discussions. I don't think that there's anything in the draft
that wasn't actually implemented first.
Except that, when it's convenient, they'll say that they can't
work on this or that because they are a "small group", lacking
massive resources.
. There are a lot of members of the committee. They all can
vote (if they attend the meetings). But each member works as
much or as little as he wants on it, and there are very, very
few people who are willing (or able) to invest a large amount of
work. It's a curious situation: the large number can (and
sometimes does) prevent useful additions, but it doesn't help
advance anything, even if they're in favor of it. Regardless of
what people want, until there's a concrete proposal, addressing
all of the issues, the committee really can't act on it.
In the end, I think they just standardize what they like/get
enough interest in.
Obviously. Those who do the work get to choose what they do
(since it's all volentary). The need for something like nullptr
was recognized as far back as 1995, but it wasn't until much
later that someone was actually willing (or able) to do the
work.
In what way you find my point of view "restricted"?
It's *your* point of view. A consensus cannot be the view of
a single person.
