Finally: The right-sized Thielert

Those of us that design computers and write software are probably less
comfortable with the software and computers than the mechanical linkages.
I've seen way to many scary things go out the door.

"Thomas Borchert" wrote in message
...
Bob,

Hardware failure modes are much better understood.

Says who? Numbers, please.

What's the reliability of software?

Not sure I understand the question.

--
Thomas Borchert (EDDH)

Soxinbox,

Those of us that design computers and write software are probably less
comfortable with the software and computers than the mechanical linkages.


Well, that's not the numbers I asked for, it's just smoke being blown. I
guess you don't have the numbers to back up your claims. Doesn't surprise
me, because I don't think the claim is valid at all. Oh, and yes, I have
designed software, too. Argument by authority still doesn't work.

--
Thomas Borchert (EDDH)

In article ,
Thomas Borchert wrote:

Hardware failure modes are much better understood.

Says who? Numbers, please.

You think you understand software failure modes?
Please enlightment me.



What's the reliability of software?

Not sure I understand the question.

exactly. Nobody agrees on what software reliability means.
On one hand software never wears out. It doesn't break.
However, it is fielded with errors. So it's already broken.
Or is it? The error doesn't manifest itself until the broken
code is executed...

And this discussion is pretty much independent of whether the
software error is a design flaw or an implementation flaw.

--
Bob Noel
Looking for a sig the
lawyers will hate

Bob,

Hardware failure modes are much better understood.

Says who? Numbers, please.

You think you understand software failure modes?
Please enlightment me.

You made a claim. You back it up.

What's the reliability of software?

Not sure I understand the question.

exactly. Nobody agrees on what software reliability means.

I guess the ton of scientific literature out there disagrees. Just one
link easily found:
http://satc.gsfc.nasa.gov/support/IS...trics_and_reli
ability.html


--
Thomas Borchert (EDDH)

Thomas Borchert wrote:
: Thielert has announced a diesel engine in the 230-HP-range, covering
: the gap between the existing 135- and 350-HP models. Great news, IMHO,
: since this is the kind of power many of the planes we fly need. First
: target for certification seems to be the 182. Time frame is 2007. Nice!

They also expanded the STC for the 135hp unit to include a larger
number of PA28. Unfortunately on the PA28-180 the MGTOW is reduced
from 2400 to 2150 lbs.
--
Aaron C.

exactly. Nobody agrees on what software reliability means.

Both Airbus and Boeing have been having trouble with flight control
software according to a Wall Street Journal article a few days ago.

Exactly what claims did I make that you want numbers for?? Did you confuse
me with another poster??
"Thomas Borchert" wrote in message
...
Soxinbox,

Those of us that design computers and write software are probably less
comfortable with the software and computers than the mechanical linkages.


Well, that's not the numbers I asked for, it's just smoke being blown. I
guess you don't have the numbers to back up your claims. Doesn't surprise
me, because I don't think the claim is valid at all. Oh, and yes, I have
designed software, too. Argument by authority still doesn't work.

--
Thomas Borchert (EDDH)

In article ,
Thomas Borchert wrote:

exactly. Nobody agrees on what software reliability means.

I guess the ton of scientific literature out there disagrees. Just one
link easily found:
http://satc.gsfc.nasa.gov/support/IS...trics_and_reli
ability.html

check out DO-178B, paragraph 12.3.4. (unfortunately, it's not available online
for free)

I've followed the FAA's software certification work since 1998. All of the
public forums and publically available work basically points to a lack of
usable software reliability models. I say public work because there is always
the possibility that an applicant has used software reliability in their
software certification work and this would be likely considered to be
proprietary.

Are you aware of any software reliability models that have been used for
aircraft certification?

Also look at the software stuff available at
http://www.faa.gov/aircraft/air_cert.../air_software/
Have I overlooked something?

--
Bob Noel
Looking for a sig the
lawyers will hate

On 2006-05-29, Bob Noel wrote:
In article ,
Thomas Borchert wrote:

It says how much a pilot wants to trust a small computer and its
software and engine actuators.

instead of a small mechanical linkage and its hardware?

Hardware failure modes are much better understood.

What's the reliability of software?

For embedded stuff, seemingly much better. Let's compare like with like:
old cars vs new cars. My first car had completely mechanical engine
components - mechanical points and condenser, vacuum advance etc. It
needed a great deal of maintenance to get any kind of reliability. It
was hard to start on cold, damp days.

In many ways, it was comparable to many aviation engines - high
maintenance and fiddly operation. Consider hot starts on even a brand
new Lycoming fuel injected engine - it needs a different procedure to a
cold start.

My current car's engine is completely electronically controlled. It
doesn't need frequent tune ups, lots of maintenance - basically, just
oil and filters. It starts just as well on a warm dry day as on a cold
damp day. It doesn't suddenly quit because something backed off and got
loose on an ignition component. It is so much more consistent than the
old completely mechanical engine as well as much more reliable. Not to
mention a great deal more fuel efficient and more powerful.

I think engine management systems for aero engines haven't come a moment
too soon rather than continuing with engines with systems from the
1940s. I'd trust an engine with electronic engine management long before
I'd trust a "traditional" aero engine. The airlines agree - all their
engines are now FADEC.

--
Yes, the Reply-To email address is valid.
Oolite-Linux: an Elite tribute: http://oolite-linux.berlios.de

"Dylan Smith" wrote in message
...
For embedded stuff, seemingly much better. Let's compare like with like:
old cars vs new cars. My first car had completely mechanical engine
components - mechanical points and condenser, vacuum advance etc. It
needed a great deal of maintenance to get any kind of reliability. It
was hard to start on cold, damp days.


We should let Honda build all engines. If they would put in access for oil
changes they could have welded the hood shut.