[SystemSafety] COTS display certification

[Peter Bernard Ladkin]

On 2016-07-27 08:37 , Peter Bishop wrote:
> In principle, any input/output function could be implemented as a very
> large lookup table. 

Not really.

Suppose machine M has memory Mem. Let RVF be a computable real-valued function RVF of real
arguments. The domain and range values of RVF can be approximated where necessary by discrete
rational numbers. The number of discrete rational numbers available is bounded by Size(M). Pick such
a set, RN. Size(RN) = Size(M).

Any look-up table T in Mem for a function F has to contain Dom(F) and also Ran(F). Assuming neither
of Dom(F) and Ran(F) is empty, then Size(Dom(F)) < Size(Mem). Which means that Dom(F) must be
smaller than RN. It follows that at least one value in RN cannot be in Dom(F).

Let F be RVF. Let a value in RN not in Dom(RVF) be V. The pair <V, RVF(V)> is not in T. Thus T does
not implement RVF.

Recall that is so for any lookup table T.

> What data validation techniques are appropriate
> to a given SIL?

https://www.amazon.de/Safety-Guidance-Initiative-Working-Group/dp/1519533578

DSIWG members Dave Banham or Mike Parsons could probably answer this specific question for us.

PBL

Prof. Peter Bernard Ladkin, Bielefeld, Germany
MoreInCommon
Je suis Charlie
Tel+msg +49 (0)521 880 7319  www.rvs-bi.de





-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 455 bytes
Desc: OpenPGP digital signature
URL: <https://lists.techfak.uni-bielefeld.de/mailman/private/systemsafety/attachments/20160727/73594e51/attachment.pgp>