ggr@hudson.UUCP (Guy Riddle) (03/21/84)
CCITT is supposedly working on standards for interconnecting electronic mail networks, to be known as the X.400 Series. The protocols involved are supposed to cover the interconnection of "public message systems" and of private ones into the public ones. Does anyone know what they are up to? === Guy Riddle == AT&T Bell Laboratories, Piscataway ===
julian@deepthot.UUCP (Julian Davies) (03/21/84)
Yes, people with the right connections can get a rather fat volume from Bell Northern research. Ian Cunningham (BNR) is the reporter for the CCITT group. The final draft of X.400-X.430 was fixed at Brighton UK in October 83, and will go to the Study Group VII this month some time, and to CCITT Plenary later this year. Earlier drafts of the documents have been circulated within IFIP Working group 6.5. The draft recommendations are numbered as follows: DR X.400 - Message handling systems: System Model-Service elements DR X.401 - " " ": Basic Service Elements and Optional User Facilities DR X.408 - " " ": Encoded Information Type Conversion Rules DR X.409 - " " ": Presentation Transfer Syntax and Notation defines an arcane variation on BNF DR X.410 - " " ": Remote Operations and Reliable Transfer Service DR X.411 - " " ": Message Transfer Layer defines the P1 and P3 protocols DR X.420 - " " ": Interpersonal Messaging User Agent Layer defines the P2 protocol DR X.430 - " " ": Access Protocol for Teletex Terminals
mark@cbosgd.UUCP (Mark Horton) (03/23/84)
For those wondering what the flavor of X.400 is, I ordered a copy and glanced through it. (The set of booklets is about 2 inches thick.) I have not had time for a detailed reading. I was trying to find out what an address looked like. Most of the publication is very broad and avoids overall examples, but one of the booklets talks about mailing addresses. What really disappointed me was that addresses are specified with binary field tags, rather than ASCII text. There is an amazing variety of attributes you can specify for a person, such as the country or sub-country-unit they are in, who they work for, their first or last name, and so on. But you can't just type it. Apparently some unspecified user interface is expected to get this information from you and assemble a binary description. It was not obvious to me how to use such a system, or how to advertise your electronic mailing address. Mark
julian@deepthot.UUCP (Julian Davies) (03/24/84)
Directory services have not yet been specified. And yes, a 'user agent' program is supposed to provide the user interface. The X.4xx recommendations are only concerned with what happens on the network, not with what kind of user interface may be built. I agree with Mark, the definitions are fairly impenetrable. I wish there were more worked examples in the definitions. The protocols are quite different from anything now happening on Usenet (or ARPAnet), and require an 8-bit wide data path for session connections. Everything is encoded in a system of 'self-describing data structures' -- where anything starts with a length code and type code(s), which is where the 8-bit wide data path is needed even if the stuff being encoded is ultimately just ASCII text strings.
jis@hocsd.UUCP (03/24/84)
OK, Here is a question that has been bothering me for a while. Since NBS is
involved in the specification of X.400 (as I gather from various sources),
is there any relationshsip at all between the X.400 and related CCITT
standards and the ARPA/RFC standards? Do the ARPA internet standards fall
within the framework of X.400? If not is anyone looking at gatewaying
problems between ARPA and ARPA-like nets to the hypothetical CCITT X.4??
net?
On the same note, it appears that CCITT is on the verge of adopting a set of
File Transfer Protocols too. How do those standards relate to ARPA FTP? Or
do they have any relation at all?
>From the ever curious keyboard of
Jishnu Mukerji
julian@deepthot.UUCP (Julian Davies) (03/27/84)
Re relationship between X.4xx (CCITT Message Handling facility) recommendations and ARPAnet conventions. Someone else ought to be answering these questions (someone who's actually on the CCITT study group, that is)! There is NO connection between the X.4xx recommendations and the ARPAnet conventions. Actually, I don't think NBS is directly involved in CCITT --they will be a member of ISO which is coordinating, and NBS staff are active in IFIP WG6.5 (Debbie Deutsch among others). NBS did draw up a draft Federal Info Processing Std on interconnection of message systems, which was influenced by a list of necessary message header fields derived from ARPA experience. See Deutsch's paper in the book "Computer Message Systems" (north-holland) which is a report of the last IFIP WG6.5 Conference. The details of the NBS encoding scheme got changed in the CCITT version. The encoding system is machine oriented, not human-oriented. The "Distributed systems research group" at UBC (University of British Columbia) is developing an implementation of a message system using the CCITT recommendations, and I believe it includes gatewaying between UUCP mail and the new protocols, but don't know the details. It isn't due for public release for a while so far as I know. Building a gateway between a CCITT system and ARPA style protoclols should be a finite task, since the same kind of information is needed on both sides; representations are different. Addressing and Naming will be a problem for quite some time to come. (Perhaps UBC will comment?) Note that the CCITT *recommendations* are just that. No coercion. Market forces will tend to make implementations that conform to them attractive, especially outside North America (maybe). Does anyone know what NBS is doing with the proposed FIPS in this area, which is now seen to be incompatible? Anyone who is really interested in this area should buy a plane ticket and get registered for the next IFIP WG6.5 Conference in Nottingham, UK, the first few days of May 1984. A notice about it was sent out in net.mail.something a few weeks ago. A book of the proceedings will be published in due course (by North-Holland, who have a monopoly on IFIP publishing so far as I can see).