Excellent; that was what I was going to start with, and if it's complete for R7RS-large, to the extent we can say that about an unfinished standard ... we need some framework for new Scheme types, or accept sdbi will be superseded at some point. Given that we want to support any relevant odd type a Scheme implementation has added, the former would be best.
Potential R7RS types are: compound objects (list-like), Maybe and Either (single-element containers), SRFI-170+ timespecs (2-element container). Less likely: mmutable cyclic lists, priority queues / heaps, doubly linked lists, edit buffers. Of course there is not going to be support for ports and procedures or anything containing them like comparators. Enums are a problem: the native ASN.1 format just sends the ordinal of the enum and depends on the ASN.1 schema, which we hates it, so we would need something more complex and I don't know what.
All covered with the above caveats.
John Cowan
http://vrici.lojban.org/~cowan xxxxxx@ccil.org "Any legal document draws most of its meaning from context. A telegram
that says 'SELL HUNDRED THOUSAND SHARES IBM SHORT' (only 190 bits in
5-bit Baudot code plus appropriate headers) is as good a legal document
as any, even sans digital signature." --me