Re: Pratt & Whitney's Embedded Software - CMM Level 3!

[Marin David Condic <condicma@bogon.pwfl.com>]

William P.Milam wrote:
> 
> Well several automotive companies are looking at the
> Mathworks tools for precisely that kind of use, I work in
> powertrain...which includes engines...which use the same
> control theory that is employed in aircraft. Only we have to
> do it cheaper. No cost overruns in teh consumer business. ;-)

Yeah, but if I don't buy a Ford and opt for a Dodge instead, you just go
persuade someone else to by the F150 truck, right? When your Uncle Sam
decides he isn't going to buy a Pratt engine, at the same time, he tells
us "But don't you dare try to sell it to anyone else!" :-)

Defense contracting is its own world and it can't operate like a
"normal" business because of that
single-customer-made-to-order-never-been-done-before kind of situation.

> 
> >(I'm also getting too old and cranky to try to once again
> > start pleading with them to try something new!)
> 
> I suspect that if you are older than I am it ain't gonna be by much.
> Being a curmudgeon is all fine and good but new things keep
> you young.
> 
Oh, I'm not really that old. I've just been working on the "curmudgeon"
part of my collection of character flaws for a long time. :-) 

Seriously: It is sometimes *very hard* to get a large collection of
people to change the way they are doing things. Especially when what you
are putting in place is brand new and not as "flawless" as what they are
already doing. It was an uphill battle & it took years of threats,
cajoling, sweet talking, bribery, and an occasional going postal of one
or more participants. Attempting to do that again with the same crowd
would not be fun.

> This is precisely what is happening in automotive, only we are
> buying not developing. At some point you find that in-house tools
> are not better and cost more. Then it's time to change. The
> cost to not change can be a great deal higher.
> 
I agree - COTS is always a better choice if you can find COTS that meets
most of your needs. At the time we undertook this system, we did first
go out and try to find what was available commercially. Nobody had
anything quite like what we wanted and the advice we got from the
vendors we talked to was "You're gonna have to make one of those
yourself..."

Given the large infrastructure investment and the long life of the
projects (and contractual obligations and a host of configuration
management issues as well as verification issues, etc. etc. ad nauseum)
there is no practical way to replace what we have now with something
commercially available. Maybe on a new project, but those don't come
along very often in our business. Remember, the F119 engine for the
Advanced Tactical Fighter was something we were developing the software
for over *ten years* ago and the ATF is just now starting to go into
production. Expect to see the ATF flying with basically the same engine
for the next 30 or more years. And we'll still be maintaining the
software under our home grown CM/Change Control/Development environment
because nobody will dare put the engine at risk by attempting to upgrade
to a new system.

MDC
-- 
Marin David Condic
Real Time & Embedded Systems, Propulsion Systems Analysis
United Technologies, Pratt & Whitney, Large Military Engines
M/S 731-95, P.O.B. 109600, West Palm Beach, FL, 33410-9600
***To reply, remove "bogon" from the domain name.***

Visit my web page at: http://www.flipag.net/mcondic