"Curtis L. Olson" wrote:
> Martin Spott wrote:
> >Erik Hofman wrote:

> >>Currently there is a problem where different platforms, different OS's 
> >>or even different compilers can get different output due to the fact 
> >>that structs are used to send data across the network. This can create 
> >>endian-problems as well as packed/unpacked struct problems.
> >
> >Oh, this paragraph is an excellent template for my plead for a
> >platform-independent FDM I/O-protocol  ;-)
> 
> What's wrong with network byte order?

Oh, it's not just the byte order. Let me try to explain my personal
attitude on this topic.

In the world of I/O-protocols I have been living for you typically had
some strange box the you'd want to get your data from plus a sheet
explaining a byte or bit mask that tells you how to interpret the data
(or how to create a adta stream to make the box understand your
intention). You didn't have to deal with the architecture of this box,
you simply had this sheet and you knew this sheet was valid ....  most
of the time. You had to deal with ongoing development (like in
FlightGear) as well, which means someone probably made a soft-/firmware
update on this box and didn't know he'd have to give you the
appropriate sheet that applies to the new firmware, but there still was
a well defined abstraction layer on the I/O channel.

When I first came to have a look at the FlightGear project my intention
was to use it as an image generator for my own project (which is
already on hold now ....) and I started looking for the correct method
to interface with FlightGear. I definitely was a bit deterred by
getting an answer in the range of "it depends" and I am convinced that
others might react similarly. Look at the AeroSim page for example:

  http://www.u-dynamics.com/aerosim/

They still recommend FlightGear-0.9.2 because they simply fear newer
versions could break the interface to their simulation library.
It would probably be good for advertizing FlightGear if FG had a "well
defined" FDM I/O interface. This doesn't have to be as sophisticated as
the Coein CIGI but at least is should be stable and platform
independent.

Just my personal view,
        Martin.
-- 
 Unix _IS_ user friendly - it's just selective about who its friends are !
--------------------------------------------------------------------------

_______________________________________________
Flightgear-devel mailing list
[EMAIL PROTECTED]
http://mail.flightgear.org/mailman/listinfo/flightgear-devel
2f585eeea02e2c79d7b1d8c4963bae2d

Reply via email to