Re: [dsfjdssdfsd] Should secure RNGs be a MUST?

[Stephen Farrell <[e-mail]>]

Well, you're brave if you're arguing to make NIST RNG specs
a MUST given recent news;-)

But seriously, I don't think its at all clear how 2119 terms
map to an implementation consideration such as the goodness
of an RNG, especially when its so hard to measure.

If there were some accepted descriptions for different kinds
of randomness requirements that might be placed on protocol
data units that could be useful, esp if we also had examples
of how to meet those requirements.

But I'm not sure if the above would be very useful really.

We may be better off just updating 4086.

However, that's what this list is for, so people can discuss
this kind of thing.

S.