GRZ027@DBNGMD21.BITNET (Peter Sylvester +49 228 8199645) (07/15/89)
> Gateways which use and don't use this mapping do not fit very well together. > > Postion 1. This was held by the majority, whose primary interest is with > "first class 822 nets" (SMTP/Internet + JNT Mail/Janet). Essentially, > the view is that the specification should not be made poorer in order to > deal with UUCP networks. > > > Position 2. Appendix A should be made manadatory, as some places need it and > it will help to avoid interworking problems (Christian Huitema and Peter > Sylvester took this line). > In the short hour before leaving to a "wonderful" 2.5 hours trip >from UCL to the airport I was trying to explain that obviously many X.400-"RFC822" gateways use at least the underscore blank mapping. There are some situations where a message coming from "RFC822" containing underscores in addresses where it would be correct to map them to a blank, and not to (u). (I took the position of the advocatus diaboli.) I think Christian's point was that he wanted to make clear that there is a problem and it is not sufficient to take a rigid position. We need at least some consensus how to phase out gateways that use undesirable mappings. I cannot go to our director and tell him: Sorry, you can no longer communicate because we are using new software. Regards Peter Sylvester
pv@SUN.COM (Peter Vanderbilt) (07/18/89)
On the Appendix A map: > There is a problem, that some nets (and in particular UUCP) need this > mapping, as they cannot support the quoted-string encoding in > addresses. It would be better solved at the RFC 822/UUCP boundary, ... Unfortunately this problem is more widespread than just UUCP. Lack of support for the Appendix A mapping would lead to serious interoperability problems. In my opinion, 987 implementations must support this mapping. Pete