And there you have it - silohed engineers managed to a standstill. There is a big problem in that there isn't the general knowledge spread across the Engineering discipline which gives rise to a great many disconnects that then needs all this documentation, control, management and checking
SW is actually very easy to understand these days with high order languages like Basic, Ada and good C.
There is a quicker way if you cut out the middle people.
On Hawk 200 we had a very clever chap who would figure out a solution, write and proved the solution in a computer program written in Basic (often over the weekend but these were not trivial problems). The solution could then be re-written into the mission computer's Ada directly - you could try reading the System Engineer's interpretation but what would be the point? - and job done. Did this several times with various weapon aiming problems and purely down to this guy's genius we had some impressive increases in accuracy.
The Germans got fed up with the delays in the Eurofighter Flight Control System SW and set up a small team that generated a fully working implementation in a few months and then proceeded to tell everybody but BAe (who I suspect they thought were the problem) how they had done it.
The trouble with having computers in the design process is you spend even longer checking and testing than just doing the old fashioned fag packet calculation and you don't get onto the real experience building of cutting metal/code and getting it working.