On Fri, 1 Sep 2006, Dirk Craeynest emailed:
"[..]
Even though some argue that low-level thread programming is not the way
to go to develop multi-core applications, [..]"
I am involved with a consortium which had used such an argument to
successfully secure funding from the European Commission for networks on
chips. The compiler vendor in the consortium unfortunately sells compilers
which are only for supposedly "ANSI-C code, possibly extended with
user-defined data types and operators using C++ classes and operator
overloading". As is typical with C vendors, actually reading what is
written by the vendor when it claims to adhere to the standard reveals
that it fails to conform to the ANSI standard: it sells a compiler which
"supports the built-in data types of the C language and allows to map them
onto the data types supported by the processor hardware. In addition,
thanks to the support of C++ classes, the compiler offers the flexibility
to introduce custom data types in the source program. For example,
fixed-point data types occurring in DSP applications can be supported
easily, without having to resort to nonstandard C language extensions (as
is required in some other compilers)."
Dirk Craeynest wrote:
"Can we do anything to make Ada more visible in that domain?
Should we respond to those papers, announcements, etc, suggesting
to also look at Ada with its long track record of successful
multi-threaded applications?"
Maybe.
"Any volunteers, who have a "good pen" to write eloquent, brief,
and to the point reactions?
[..]"
Not I.
">>> Looking for Ada-related jobs, reachable from Leuven, Belgium <<<"
Good luck with that search as well.
|