[sv-ec] proposal for 261 and 262

It was difficult to write a proposal for 261 that was not affected by 262,
so I wrote a combined proposal that I attached as a file to 261.

It has two possible variants for the changed text. One always does the
evaluation of the weight expressions as unsigned, which effectively means
that all the intermediate values and operations are unsigned also. The
other does evaluation of the weight expressions using context-determined
signedness, like case statements, and then converts the results to unsigned.
This is different if all the expressions are signed, since the intermediate
values and operations will be signed instead.

The first variant is simpler to describe, but makes it difficult to do
signed computations in a weight expression (though there are ways to do it).
The second is harder to describe, but may give more natural-seeming results.

Steven Sharp
sharp@cadence.com
Received on Mon Nov 8 14:38:30 2004