XML version of joined IDL
From: Sigi Reich (sr_at_ecs.soton.ac.uk)
Date: Fri 27 Nov 1998 - 10:08:10 CST
it's Friday afternoon again and we would like to update you on our latest
achievements in interoperability in navigational hypertext. Dave and I have
been working on a joined XML representation that is based on
- the Danish proposal for the OHSWG meeting
- findings of discussions on this list
- our own stuff
We have decided to separate out the dynamic services and "advanced"
collaboration stuff leaving a pure navigational interface (the necessary
hooks such as session info, computationIDs, etc. are in the DTD).
The resulting DTD can be obtained as ohpnav27Nov1998.xml from
A few questions remain which are included below.
Any feedback warmly welcome!!!
Dave & Sigi
These are the comments from the English jury (specifically concerning the
Danish "entry" ;-))
- interface Id: what exactly are OS, ENT and LOCAL
- interface AbstractObject: PSPecIds should be in HMObj (according to data
- interface AbstractObjectFlags: most flags can be ommitted in the XML
representation because the simple fact that an element type is present
means that it is relevant. We appreciate however, its value in the IDL
spec. For retrieval we have added a flag.
- interface AnchorEndpointSet: what is this for?
- all sets are done is * element types or IDSET in the XML representation
- we felt that subscriptions were part of the advanced (i.e. domain
specific) collaboration stuff, is this correct? because of this we left
them out of the XML representation of OHP-Nav
- what xactly is the difference between xactId and Id in e.g. createContext?
- what is the difference between link and context (from a data model point
of view)? Both have the same structure, the only difference is in their
type. Perhaps we should restrict the IDSet of Link to EndpointIDs?
- we could not find a function / interface for "sessionStateModified"
notification in the client.idl
- in the data model: wouldn't we still need an IDLOC? How would one link
between say an endpoint in a document and a PSPEC? We discussed this at the
working group meeting in Southampton and the argument was that a locspec
with an ID would be modeled as the "reference" attribute of the locspec.
Now, the problem is that there is no such thing as a locspec (it is an
abstract type): it is EITHER an NaLoc OR an AXISLOC. Hence we would need an
IDLOC (as we have proposed earlier already).
- linkfollowfromlocspec: how and where is "direction" used? And: we thought
that each message could be sent with a set of contexts (contextIDs). Hence
we do not need to specify a contextIDs parameter for this function. Because
we decided at the OHSWG meeting that it was undefined when to execute
computations we left the compID out.
- linkCreateWithEndpoints: shouldn't this be either a createLink or a
- what about functions such as displayDataRef, closeNode, etc. Are these
still considered useful?
BTW: The dtd has been parsed with Internet Explorer 5.0 and therefore
should be correctly formatted.
Multimedia Research Group
Department of Electronics and Computer Science, Bldg. 59
University of Southampton, Southampton S017 1BJ, UK
phone +44 (0)1703 59 6505 fax +44 (0) 1703 59 2865
email sr_at_ecs.soton.ac.uk http://www.mmrg.ecs.soton.ac.uk/
This archive was generated by hypermail 2.1.5
: Tue 13 Aug 2002 - 07:21:09 CDT