( ESNUG 338 Item 9 ) --------------------------------------------- [12/3/99]

Subject: ( ESNUG 335 #9 )  OK, C/C++ HW Design Won't Work; Why Not SLDL?

> My belief is that trying to embed everything in one language is just part
> of the eternal quest for the silver bullet.  Fred Brooks wrote an IEEE
> Computer article on silver bullets (with a cool werewolf picture) many
> years ago.  One obvious issue that is not mentioned is that the software
> and hardware developers are rarely the same person.  We don't expect the
> circuit and layout designers to use the same tools.  What is needed are
> tools appropriate to each problem domain, but which enable integration,
> co-design, co-simulation, co-verification, etc.  One language for
> everything is another way of saying lowest common denominator.
>
>     - Duncan M. (Hank) Walker
>       Texas A&M University                  College Station, TX


From: Steve Schulz <ses@ti.com>

Hi John,

The result of 5 industry workshops (each one comprised of 30-50 leading
experts in the field worldwide) arrived at the same conclusion as Duncan
Walker states above, based on quite rigorous technical and usability
analysis.  A similar analogy could be made with software systems.  One
could write a full GUI-driven application entirely in assembly (or
low-level C), or one could write a device driver out of perl and a few
extra foreign interface bindings.  Yet we use a blend of assembly, C, C++,
Perl, Tcl/Tk, MS-Excel Visual Basic macros, etc.  The reason?  It's the old
adage: "use the right tool for the right job".  Once we truly understand
the greater complexity of the "100M-xtrs-and-embedded-sw-on-0.1u" design
task, it is clearer to see why we should not expect C or C++ to solve the
entire problem standalone.

That is not to say that C/C++/Java won't have a useful role.  The trick is
that we need to find ways to describe, analyze, and "prototype" all aspects
of these systems before they are actually committed to expensive silicon.
A means to integrate the views, and different syntaxes, is needed, and must
be interoperable for the industry to leverage from any of them.  The System
Level Design Language (SLDL) effort is now at work defining such an
integration system.  I expect that we will see (one or more variations on
C/C++ as an integral part of the SLDL environment with Rosetta, but it is
not realistic to expect any single software language to "be all things".
(see http://www.inmet.com/SLDL for more background.)

    - Steve Schulz
      SLDL Initiative Founder
      Texas Instruments                         Dallas, TX