All,
I agree with almost everything in Paul's proposal. Our (VDA) v3 X.509
Certificate implementation uses a 1984-based ASN.1 Toolkit. We successfully
used the strategy proposed by Paul to implement UniversalString. We defined
UniversalString as [UNIVERSAL 28] IMPLICIT OCTET STRING. We have proven
that our implementation of UniversalString is interoperable with several
other implementations, one of which was using the "modern" OSS ASN.1
ToolKit. This is the same strategy that Paul is proposing for UTF8String.
The part of Paul's proposal that I question is the inclusion of languageTag
in ESSFreeText. I would like to avoid forcing the receiving software to
check for two sources of languageTags. It would be much cleaner if we could
always use embedded language tags in the UTF8String. I understand Paul's
concern that the concept of embedding language tags in UTF-8 strings is
immature. However, if we believe that the concept will be nailed down in
the next few weeks, then I believe that it is worth the risk to simply use
UTF8String. If embedded language tags is a risky concept that may get
thrown out altogether, then I believe that we should stick with Paul's
proposal.
================================
John Pawling
jsp(_at_)jgvandyke(_dot_)com
J.G. Van Dyke & Associates, Inc.
================================