<< So Rick, the bottom line is; the  Windows 95 OS cannot be certified by the
FAA.

Is that a true statement? If not why not?
If it is the lack of source code would it be easier to certify Linux?

I believe that Windows 95 could be certified (with a lot of effort) to
DO-178B level D.  This is what Avidyne has done with NT.  BTW, it would be a
specific version and build of 95 that would be certified.  Any
changes/updating would require recertification of the changes and
verification that unchanged functions are truly unchanged.

With the availability of source code, Linux could theoretically be certified
to a higher software level, again, with a lot of effort.

I am wondering why more people are not offering PC compatible cockpit
computers in every sense of the word. I understand the need to make more
money by selling and locking customers into  proprietary software.
Is that the reason or is it the refusal of the FAA to certify Windows 95? >>

Paul, it's not a conspiracy, it's a market and certification reality.  When
Avidyne started many years ago, their idea was to produce a certified cockpit
PC and let others produce certified software.  You could even download
developers kits from their website.  Well, producing certified software is
not a trivial task.  Nobody ever wrote software for their system.  There are
several reasons for this:

1.  Most pilots, especially those with certified planes, don't want to fiddle
with their cockpit equipment like they would do with a home PC.  They want a
turnkey solution, installed by an avionics shop, that works.  They usually
won't even read manuals.  The above goes double or triple for OEMs.

2.  The FAA has stringent requirements for version controls, manufacturing
controls, and field-loading controls for certified systems.  A certified
cockpit PC would require certified hardware and certified software.  These
components would have to be certified to work together and STCd for a
particular aircraft with FMEA's, unique HIRF testing, etc.  In multiple
application systems, each application must likewise be certified to work with
or shown not to affect all other applications.  It is a tremendous burden
even when one company is handling all aspects (hardware and software).
Spreading the responsibility among more than one company, absent a very close
working relationship, would be nearly impossible.  Note that Avidyne and
Avro-Tec are cooperating in this manner with their Lancair product on
low-criticality applications (you don't see any engine monitoring or primary
flight displays on their screens -- mainly due to their use of NT).  I don't
know of any others.  Also, the Avidyne/Avro-Tec partnership is aimed at
delivering a turnkey certified system.

3.  Each cockpit PC has a unique control planform.  It has been shown over
and over again that pointing devices and touchscreens simply don't work that
well in the flight environment.  Large, finger-size, high-quality buttons
with hand-steadying structure available is the standard for aircraft systems.
 And remember that the light-GA cockpit is a much bumpier place than the
airliner or fighter cockpit.  The software interface must work with the
hardware layout.

Of course, in the experimental world, all of this goes out the window (no pun
intended).  There are many makers of general purpose PCs intended for cockpit
use that can be used in experimentals.  I think that the FreeFlight website
used to have links to them.  Preston Peavey and Avro-Tec come to mind.  I
don't know how successful these companies have been, given the limited
market.  I sure don't see many out there on the flight lines (i.e. I see more
rotaries :-)).

Thanks for your insight on this Rick. Mind if I put it all on the web site?

Paul Lamar

 
The Aircraft Rotary Engine Newsletter.              Powered by Linux.

http://home.earthlink.net/~rotaryeng/            http://www.linux.org