RE: [rohc] Role and approach of the ROHC formal notation

[Carsten Bormann <cabo@tzi.org>]

> The EPIC-lite approach is more efficient than the 3095 approach,
> so I don't know why Carsten thinks we need separate discriminators.

I'm not so interested in the separation, just in a way to control the
combinatorial explosion.  It would then become a matter of design to
control which indicators combine to form one discriminator (which is then
optimized based on the known probabilities of the possible values) and
which ones stay separate to reduce combinatorial explosion.  The latter
may be desirable when the designer knows that the case in which the
indicators are used is unlikely enough, and also when it is somehow clear
the values being separated from each other are sufficiently orthogonal
that there is little gain in the optimization.

All this is, of course, just a set of second-order considerations.

Gruesse, Carsten
_______________________________________________
Rohc mailing list
Rohc@ietf.org
https://www1.ietf.org/mailman/listinfo/rohc