Re: Ada to C Translator

[Dr Adrian Wrigley <amtw@linuxchip.demon.co.uk>]

Ted Dennison wrote:
...
> I wonder if other walks of life have this problem... Do civil engineers
> rip down a perfectly funtional and stable bridge and rebuild it with the
> same carrying capcity using different materials?
...

Yes!  Here in Cambridge (England), civil engineers pulled down a very
well constructed bridge (The Victoria Avenue Bridge), made of stone
and iron, to replace it with something visually similar of steel.
The cost (IIRC) was several million pounds, and a lot of inconvenience.

The reason was that the original bridge could not be proved to be strong
enough for the modern traffic load, because inspection of key structures
was not possible, and the plans were not available.

Only once the project was well under way, and the bridge was in pieces
was it possible to tell that the work was unnecessary.  The work
completed as planned. (corrections to my memory of this welcome!)

To switch back to the underlying topic...

I agree with those who say that Ada to C porting of source code
is of extremely limited value.

If Ada were prohibited for the project, you might find it useful to
generate C source files as part of a multi-stage compilation process,
and then tweak or replace certain C files as necessary.

I imagine that such an approach might be useful with 'end of life'
applications, when someone suddenly offers lots of money for it to run on
a machine with no Ada compiler (Alpha Linux? Macintosh? ARM?), and a
short timescale.  Or maybe the customer demands C source in the contract.
Or maybe the management doesn't have a sufficient supply of willing
Ada programmers to maintain the project, but doesn't like the thought
of a total rewrite.

just a thought...
--
Adrian Wrigley