Prochnow said:
The following is an interesting read on the 45...
http://www.theregister.co.uk/2004/09/06/ams_goes_windows_for_warships/
Being in the RNR my day job is IT support. I do have an overly nerdy interest in the 45s Command Operating systems. I think I can say with some conviction (12 years and a degree in information systems) that anyone with any ounce if IT support knowladge knows UNIX is far superior than any MS operating system in terms of performance, reliability, security and robustness. So regardless of what any "independent survey" says I'm frankly gob smacked that the RN would allow the Command Operating System to be Windows based.
As a professional safety engineer (not health and safety, but making sure things work correctly and safely) I don't have any pre-conceived reservations about which OS should or should not be used; however as a professional safety engineer I would expect to see a reasoned argued justification, in the form of a safety case, that failures within the OS cannot and do not affect safety (or functionality, reliability, availability or any other "-ility"). Sadly, from first hand experience of working with the developer, I don't think such an argument will be presented - it will more likely be "here it is - take it or leave it!".
Looking slightly deeper into the "why" of the switch; from reading the article, it appears that the drive to Windows has not been driven by technical requirements arising from the designers and/or customers, but is instead a result of AMS' (now BAeSystems) use of a sub-contract IT department's desire to remove all things "non-Windows". FFS, since when did design decisions get taken by sub-contract service suppliers?
This decision reflects two major problems with today's defence engineering approach:
1. There is a major ignorance within MessyBeast's organisation about the distinction between the systems that it is producing to supply to its customers and its own internal IT systems - while it may be reasonable to mandate MS for the latter on grounds of commercial efficiency, such restrictions must not be applied to the former without the consent of the design authorities.
2. With the virtual monopoly that MessyBeast has created for itself, I doubt whether the DPA has the necessary power, nor indeed the motivation, to insist that a sound engineering decision is instead of deferring to a commercial decision taken by a sub-contractor.
Prochnow said:
Quoted in the above URL “Anyone with elementary knowledge of computer science can see that Microsoft Windows, as described here by Gates, is inherently insecure by design.†is – for me at least – fascinating.
Security is not my concern, I'm more worried, nay scared, of the fundamental non-determinism of Windows (i.e. no-one, not even the MS designers, can predict with certainty what it will do) and as such it should not be allowed to be used within any system interfacing with a weapons system, especially not when such a system could support autonomous engagment.