array 
Today I've learnt first hand of an enterprise with an IT staff of ~200 people stuck on a Java 6 (from... 2006!) codebase. Correct me if I'm wrong, but I guess they must be on their own, because it's been a while since it went EOL. That makes one think again on how difficult must be to migrate such huge codebases... And how difficult must be to patch it to keep the engines running on a daily basis. Sigh, computers were supposed to make everybody's life easier, but. :P
The Shell of a Fish Finger@array theres an awful lot of space stuff still running Java 6 (and earlier). You do not rewrite and recertify mission critical code unless you have to and someone else is paying.
@fishfinger Yes, I believe that may be true. But it still sounds a bit... Scary, I mean, running officially deprecated code in mission critical production apps... But that's how it is I guess. ;)
@array Their view would be that its a lot scarier running code that Oracle says is good but hasn’t been running for a decade.
The testing processes are insane though: aside from everything else you’ll spend about a month (even for relatively innocuous subsystem) stepping through every possible state the machine could be in and manually checking that it does what the spec says in front of lawyers and engineers from higher up subcontractors and ESA: any deviation gets the lawyers involved. Its brutal.
@fishfinger Sounds like that indeed. :) But I understand that sometimes a minimal bug can be very costly, just a couple hours downtime can translate to very big losses for some medium to big players, so I understand testing is very critical. ;)
@array It’s more that you don’t want to delay a launch or kill an astronaut. Things like the Mars missions have their launches scheduled decades in advance when the planets literally align. If you miss the launch the (contractual) costs are high.
And astronauts are just too valuable to risk. They are idolised.
@fishfinger Makes sense. :)