"autocode" vs "intermediate code"

[Stephen Leake <stephen_leake@stephe-leake.org>]

R. B. Love <rblove@airmail.net> writes:

> On 2007-02-18 19:13:47 -0600, Ed Falis <falis@verizon.net> said:
>
>> Jeffrey Creem wrote:
>>
>>> Do you have first hand knowledge that ISS has a lot of Ada.
>> Yes. Or at least I can't believe they rewrote all the code they wrote
>> (often using code-generation tools) for the MDM's.
>
> Regarding code generation, I know that MatrixX was used for part of
> the GNC MDMs.  The CMG control logic was designed by Huntington Beach
> and autocode was created.  The navigation and monitoring code was
> written by hand in Houston.  Guess which one is readable?

I assume the Ada original source for the navigation and monitoring
code is more readable than the Ada intermediate code for the GNC. But
that's like saying the assembly intermediate code for the navigation
is not as readable as the assembly original source for some core
floating point function.

Using the term "code generation" instead of "compiled to intermediate
code" is misleading. It is more appropriate to say the GNC code was
written in MatrixX, and Ada was just an intermediate language on the
way to machine code.

The proper question is whether the Ada original source for the
navigation is more readable than the MatrixX original source for the
GNC. For myself, I find most MatrixX (and Matlab Simulink) source to
be hard to read, so I would probably find the Ada code more readable.
But some people prefer the graphical languages of MatrixX and
Simulink; I assume that's why they used it.

A related question is whether you can still buy support for the Ada
compiler and the MatrixX compiler. MatrixX went out of business; is
the Ada vendor used for ISS still in business?

And how many MatrixX or Simulink programmers can you find on the
market? I bet it's far fewer than Ada programmers. Yet no manager
would hesitate to allow Simulink source on their project, because it's
the "in" thing.

-- 
-- Stephe