Dear SG16ers,

This is probably a bit late, but I was reading through Annex G and noted that there were a few macros used (principally for CryptoH323Token). These make it much harder to write a compiler. If it is still an intention to have Annex G widely accepted by making the message encoding easier, then removing the macros would help greatly to this end.

I can't find a recent copy of H.235, but if memory serves me right, I feel that CryptoH323Token could be written out long hand with no loss of functionality, although the definition wouldn't be so brief.

For signing messages I would propose allowing a message to contain more than one separately encoded ASN.1 chunk. The main bit would be the message, either raw or encrypted. The other optional bit would be a security definition, including signatures, and possibly key ids etc. There may also be a need for a third un-signed or plain text ASN.1 chunk to allow gatekeepers and other entities to modify message contents without invalidating any signatures.

Be latedly yours,

Pete

============================================= Pete Cordell pete.cordell@btinternet.com =============================================