Re: [Gnu-arch-users] Nit

[Michael Teichgräber]

address@hidden writes:

> While this is an obvious troll..

Certainly.

> That article from 1994 is a fud article that many have quoted as proof
> while it just points out some problems in someone's code.

As a proof for what? I didn't quote it as a proof, just as an
article. "some problems" might turn out to be rather complicated.

I found said article some years ago, when I had reasons to look a bit
into such problems. C++ and related technologies just presented
themselves to me not as more easy, more secure, more reusable and
reducing "time-to-market" as claimed, in the way they were used and
pushed in an automotive company where I worked -- compared e.g. to
well-written, i.e. readable, portable, reusable, careful written C
code, also to be found in a niche in that company; and it wasn't fun
to work there and see some major projects fail because of things like
"strategic partnership" with M$, Windows CE, all-apps-in-one-process,
C++, XML, COM, PVCS ... coming along with the--in my eyes--real and
damaging fud. This means, there were reasons to think about which of
the tools/languages we were forced to work with were really better
than others, which we would choose if we would decide ourselves, and
whether we simply should do so.

This doesn't at all speak against C++ and exceptions as such -- the
above is only a sketch of some of the circumstances (where I got in
touch and didn't really make friends with these things) I remembered
when I followed the recent postings, and I thought: 'There was this
exception article'. To be exact, I didn't really remember or think to
post something until I saw your mail with the embedded URL. (Perhaps
there is an association in my brain combining exceptions,
URL-to-an-article, and sending-a-mail, since once I sent a mail
containing the URL to that article to a collegue.)

> I find it amazing that nearly 10 years after date people still believe
> this article!

That's not much. For instance, sometimes I find it amazing that people
(including me) use computers at all.

> Anyway; see the 4th parag in h2 of
> http://www.boost.org/more/generic_exception_safety.html
> for a debunk.

I didn't think that a solution of that problem may not be possible,
when I posted, so I had no reason to take this as a proof for
something. The "debunk" says, he "unfortunately fails to present a
solution to his problem". Did he fail? He didn't try, and perhaps
didn't want to solve it.

So far I didn't get the impression that Cargill was just wrong, or
simply an anti-C++/exceptions troll. As I understand it, today most
C++ programmers needn't read, or believe in, his article, as now there
are many exception-safe components/examples they can take advantage
of.

-- 
Michael