[SystemSafety] COTS display certification

[Peter Bernard Ladkin]

On 2016-07-29 11:13 , Peter Bishop wrote:
> PS Your impossibility argument works if you assume Size (RVN) = size(M),
> but you could could choose size(RVN) to be *less than* size(M), i.e. use
> a coarser discretisation to ensure the lookup table fits the available
> space.

I would say no.

If you are mimicking a real-time function then you need all available (approximate) values, as
accurately as they are available.

Alternatively, you could decide to approximate your real-time function F by a discrete, finite
function FF. Then you can freely choose size(FF) < size(M). But then you are not representing F by a
lookup table, you are implementing FF as a lookup table, and FF is a different function from F.

That has consequences. If you are trying to solve PDEs, then attempted solutions might be modelled
or errors might take over at any given set of inputs for which you are attempting to solve, and you
need to decide between those two cases, and muck around with the inputs to manage the error
behaviour. Whereas, if you take the grid as given, the discrete operations give you what they give,
and there is no need to decide anything. And you are not solving a PDE.

I don't see there is any way even in principle to implement a PDE solver as a gigantic lookup table.

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/20160729/f4e6ba4b/attachment.pgp>