General design

[Rodrigo Damazio]

Thought I'd start this thread to discuss what the open-source avionics should be like (without getting into implementation details yet).

My thoughts so far:

• (Obviously) it should be open - both hardware and software
• It should have the reliability of the current avionics systems, without losing the flexibility of an open system (e.g. by isolating reliable and potentially unreliable software/hardware)
• It should be completely modular so each person can pick just what he wants to install (both hardware and software)
• It should have strict compatibility standards for both hardware and software, such that any module built for the system is guaranteed to work in a plug-and-play fashion with any other modules
• It doesn't need to be super-smart in its core, but should allow software developers to implement additional smart functionality
• Security should be a concern - pieces of software should have the freedom to control airplane functions to some extent (e.g. for automation), but the user should be made aware of what each piece is doing and should be able to very quickly disable/override it if needed
• It *should* support some form of Internet connectivity, even if only while on the ground. This can be used for anything from auto-uploading performance and GPS data to downloading software updates or new databases.
• It needs databases (both VFR and IFR) - it may be out of our scope to build those as well, so this needs further thought.
• It should support some level of compatibility with existing devices, such as ARINC standards
• Having certified implementations/installations is a nice-to-have feature for the (far) future - it's expensive and takes a ton of work, IMHO we should start with experimental only

Thoughts?

[Jason Gillman Jr.]

Good points. I really like taking a modular approach to this, as it allows not only flexibility, but security, in the sense of being able to fence a component  if needed, as well.

I'm thinking that internet connectivity might be a good candidate as being a core module.

As for the databases - it would be nice if the public could still access DAFIF data. However, a related sibling project could be started as a result of this.

Another core module should probably be developed to handle and display geo-referenced imagery.