Re: New Language Design: (Cray, Sun prep radical software models for petaflops systems)

[Dan Nagle <dnagle@erols.com>]

Hello,

On 26 Nov 2003 15:06:21 +0100, Florian Weimer <fw@deneb.enyo.de>
wrote:

>"Warren W. Gay VE3WWG" <ve3wwg@cogeco.ca> writes:
>
>> I find it interesting that they classified "Ada"
>> as a failure, and are looking at ways to extend
>> "Fortran" (of all things) and C++.
>
>How many users of supercomputers write their programs in Ada?  How
>many vendors sell Ada implementations for vector processors?

Well, Cray used to have an Ada (83, IIRC).

They also had a Lisp (Franz).

The idea was that the Lisp portion of a program would manage
the missile simulation, the Ada portion would direct the missile,
and the Fortran portion would report stresses, airflow etc.

The interesting part was that Lisp could change, e.g., the definition
of square root, in the middle of execution.  But that had better
not change the Fortran & Ada idea of arctangent!  It also probably
added some, ah, features, to the linker.  ;-)

It was a great idea, it just didn't work.  :-(

-- 
Cheers!

Dan Nagle
Purple Sage Computing Solutions, Inc.