The US military had a project to unify all their hardware radios into a mega SDR (software defined radio) project called JTRS. It was a massive debacle and didn't deliver very much. One of their bad decisions was to organise it around CORBA. I was amazed when I read that - CORBA did not solve a critical issue for them; it was just a big distraction. You can think of SDR as being a bit like ffmpeg or gstreamer - a collection of DSP modules. Plugability is a useful feature of those codebases, but the critical requirement is efficient data flow. IIRC this all happened somewhat after CORBA fell out of fashion in the software world too
sillywalk · 6h ago
Ars Technica has a couple of good articles about JTRS from 2012[0] and 2018[1].
EDIT: Link from the 1st article to a report by David Axe @ The Center for Public Integrity[2].
What they should have done is a smaller project to roll up all the old legacy radios into SDR, leaving the new ones as custom silicon. That would have stood a decent chance of working, and then they could go for the more ambitious goals later .
Problem is that it would not have got the people working on it into the exciting (and higher revenue) new radios.
p_l · 5h ago
From my (limited) exposure, CORBA might have been the least problematic decision in the whole project
EDIT: Link from the 1st article to a report by David Axe @ The Center for Public Integrity[2].
[0] https://arstechnica.com/information-technology/2012/06/how-t...
[1] https://arstechnica.com/information-technology/2018/03/the-a...
[2] https://publicintegrity.org/national-security/failure-to-com...
What they should have done is a smaller project to roll up all the old legacy radios into SDR, leaving the new ones as custom silicon. That would have stood a decent chance of working, and then they could go for the more ambitious goals later . Problem is that it would not have got the people working on it into the exciting (and higher revenue) new radios.