Shai,
that's a good suggestion. When a standard has a "free parameter", it
is best to state which values of the parameter are mandatory to
implement. In this case, I expect that we'd want to say something
like "Implementations may support multiple block lengths, but need
not do so in order to conform to this specification."
David
On Jan 17, 2006, at 11:52 AM, Shai Halevi wrote:
> The point is that extending the transform to handle sectors of other
> sizes is just that - an extension. [...]
Which reminds me: since the standard supports mutiple sector
lengths, we
need to add some language to say that a compliant implementation
need not
support arbitrary length (e.g., an implementation that only supports
512-byte sectors should be able to claim compliance with the
standard).
-- Shai