Sunday 6 March 2011

j661 Server ln Mac OS X

I've just bought a Mac Book to do some development on iOS (I hqve an iPhone and pkan to buy an iPad 2 when it will be available in France). First let me say that there is almost no configuration necessary for being able to use it. And despite all rants about Apple support for Java, Java 6.0 JDK is already installed ;)

Well while I downloaded XCode, I wanted to check if the j661 Server could run on Mac OS X out of the box. Well it does. Which is only natural, because it is one of the reason why Java is around there, isn't it?

There's nothing much to say, if you own a Mac, just download the last binaries fron the project page. double-click on arincEditor.jar, and enjoy ;)

4 comments:

Unknown said...

Hi, I've got one question regarding Java - is Java used on real avionic devices too? Reason I'm asking is because I'm going through a Do178 book and they mention that determinism is such a huge issue in avionics certification - which is why they recommend Ada, C or C++ (and even that without stuff like virtual functions). So can still Java be used for a CDS?

Unknown said...

Hi, I've got one question regarding Java - are you guys using Java on real avionics hardware? Or is this for rapid prototyping only?

Unknown said...

Sorry for the double post. The first post was meant to be a preview.

mithrandir said...

Hello,

Real-time Java is used on real avionic devices. For example I heard it was used on American UAVs.

But we are using j661 only for prototyping and environment models. However we still need these prototyping models to perform in a " soft real-time" simulation, so at roughly the same rate as the "real" real-time system would do.

As for determinism, you are right, but I think that IMO there are far more problems to handle in a DO-178 process than only the language used.

However, I think that the main problem with using Java in a DO-178 context would / will deal with the dynamic memory allocation, more than only the relative non-deterministic time. There are usually a lot of non strictly deterministic behavior in certified avionics systems. What would never be accepted is not the fact that something can happen at one time, "or shortly after this time", but to be able to prove that it will happen in a "reasonable" time-scale. What "reasonable" means will depend on the function and the context. Sometimes being late of 10 ms will be too much, sometimes a delay of 0 to 320 ms, or even more, will be ok.