[saag] Relative OIDs are the simpler form of OID (Re: A si mpler form of OID)

> The one thing I really, really like from ASN.1 that I haven't seen an
> equivalent of in XML or JSON is the OID, an extensible identifier
> format whose length is self describing. [...]

FYI the _value_ part of a DER-encoded OID does not indicate how many
sub-identifiers there are in the value.  The only lengths that are
self-described are the lengths of each sub-identifier, not of the whole
OID.

Anyways, what about using RELATIVE-OID?

A relative OID is basically the same as an OID, but the first sub-
identifier does not encode the first two arcs of an OID, and also a
relative OID is meant to be relative to an [absolute] OID.

You could then say that for any given relative OID slot in any given
protocol the missing prefix is as specified in the protocol's spec, and
you could have a leading arc value for any such relative OID namespace
that says all arcs below are private use or to be derived from hashing
as you suggested.

The point is that you could reuse the OID/RELATIVE-OID encoding, retain
the OID registries (such as they are or aren't, though here we could
have IANA run each sub-namespace), and get what you wanted but without
inventing a new thing nor a new encoding.

Neither OIDs nor relative OIDs are self-describing as to length -- you
have to convey the length of the encoded OID separately.  You could, I
suppose, apply your odd/even idea to relative OIDs to get something like
what you wanted so that these identifiers are self-describing as to
length, or maybe encode the number of sub-identifiers in the first one
(or include that count outside the relative OID), or require a fixed
number of sub-identifiers.

I can imagine an OER-style protocol where knowing that a value is
self-describing as to length the length is omitted, sure, but generally
OER-style encoding rules omit lengths when they can because of value
constraints that fix the value size, and TLV protocols never omit
lengths (because they're not as awesome as OER-style encoding rules).
Thus I expect that you probably don't need any such value to be
self-describing as to length, thus relative OIDs should fit the bill.

Nico
-- 

_______________________________________________
saag mailing list -- [email protected]
To unsubscribe send an email to [email protected]