A set-key attribute for symmetric-key packages
Summary: Needs a YES. Has a DISCUSS.
Alexey Melnikov Discuss
Discuss (2011-03-07 for -)
I have small issue I would like to discuss before recommending approval of this document: 2. The set-key attribute As above, this structure may be expanded at a later date with additional types. Implementations SHOULD gracefully handle values and types which they do not recognize. How can the SHOULD be satisfied in an interoperable way? By ignoring them? How is extensibility going to work in a general case? Handling of a union versa handling of an intersection are going to produce quite different results.