Perry,
Your point about DNS security extensions is a good one in so far as
it shows at least one example of a (still to be deployed) Interent protocol
where signatures will be employed. However, DNS security deals with
existing data structures in the DNS database, and it would not have been
appropriate to use ASN.1 for those non-ASN.1 data structures. I'm also not
convinced that the range of data types in DNS records is nearly as
extensive or as complex as those that we will deal with in the PKI
environment. Also, the PKI is dealing with X.509 certificates, which
already embody ASN.1 syntax, and so a developer has already bought into
whatever costs come with ASN.1 in order to process certificates and/or data
that will go into fields in these certificates. On those bases, I think
ASN.1 is an appropriate default for protocol syntax for PKI messages.
Still, I agree that we should be open to alternatives. The WG could develop
a set of criteria for the syntax specification and encoding for messages
and evaluate alternatives including ASN.1. I worry that this might be
something of a distraction from our main objectives, but if there is
sufficient interest in the WG, we can pursue this apporoach to evaluating
alternatives.
Steve