 | |
SWAD-E
Calendaring
workshop
9 Oct 2002 | | |
 Intro
 Libby, SWAD-E calandar
From XHTML profile, using XSLT
Damian Steer, iCalendar standard, Mac stuff
Libby working on RDF encoding of info.
iCalendar Large, complex, non-extensible
Information mixing
CoDepiction
DanC travel tools
Why RDF? Merging works for free.
cwm output string rules: x outputstring "y" (could use for N3Report)
Rules for itinery checking; e.g. make sure return flight after end of meeting. Lots of rules!
cwm has HTTP cache -- part of log: semantics builtin
"If they move the airport, I don't have to recompile my code"
iCal interoperability "evolution" is case sensitive. Dan also said somewhere "unused
data rots" => don't model more than you need.
Extensibility: command-line namespace for X- properties (hmmm... what if several independent X-
used?)
Synchronization hard problem: e.g. correcting mistakes
DanC has perl/python scripts to convert between RDF & iCalendar (enough for "evolution")
-- baling wire & ...
Discussion of vocabularies to use: DanC suggests one schema for each native format:
- syntax translation for native <-> RDF
- rules for mapping between schema
This strategy seems to work.
DanBri, Mozilla calendaring
Mass-market calendaring is available
Calendars for many different purposes; grassroots; things jumbled up; strength of RDF
- e.g. movie schedule
- webcal: URI scheme; subscribe to calendar; polls subscriptions on startup??
- published calendars can be loaded into Mozilla calendar display
Mozilla doesn't use RDF, but is open source. Starting to use RDF machinery in Mozilla.
Not interested in quest for perfect representation (how does that square with formal inference?)
Not yet linked times to places. iCalandar has text field location, also long/lat.
(Mozilla has built-in XSLT engine, and directory of converters)
"Hack it together first, then the hacks drop away as the web matures"
Cwm relation flattening, not like Prolog - more like SemaFor frames (thinks: use my rule formats for
this?)
Recurrence of "a little bit of inference goes a long way"
 Terry Payne? (ex CMU), Outlook integration
Insert RDF events into outlook
Security problems with calendar agent - need better trust modelling, or finer grained authorization?
Retsina group - various agents to deal with different kinds of information; chaining though agents. (Except agent was Google-scraper, they changed their format.)
Agent description language, like WSDL/UDDI ideas?
Events can be captyured, filtered, chained, and dumped into outlook. Also with URIs so that details
can be periodically checked.
Importance of provenance to back out data from corrupt source
Ontology translation service; use of service discovery as part of chaining infrastructure?
Repeating events... difficult to model...
Greg, iCalendar
Also, Orlando, logical mpdel for time. Relates to iCalendar.
iCalendar: what's happening?
IETF WG CalSCH, no meeting in Yokohama; process of creating new version is "dead", working
on access protocol. WG participation has narrowed.
XML has created much confusion about what to do with this and other activities.
See:
http://lists.w3.org/Archives/Public/www-rdf-calendar/2001May/0043.html
- (Notes relation vCard, iCalandar, MIME directory, ...)
Average number active lines/iCal object: 4-5
...
Orlando model, concentrate on temporal aspects of meetings - points out difficult issues
Calendar vs periodic (e.g. 30 days interval vs 30th day of each month)
And other issues...
MikeD points out .. is new DAML-time working group (!)
...
reoccur // v.intr. (reoccurred, reoccurring)
occur again or habitually.
reoccurrence // n
...
recur // v.intr. (recurred, recurring)
1 occur again; be repeated.
2 (of a thought, idea, etc.) come back to one's mind.
3 (foll. by to) go back in thought or speech.
[Latin recurrere recurs- (as re-, currere ‘run’)]
Per, SKiCal
Movie theatre schedule use-case
What happens in a city? Tourist events; 100 each night. Need to have structure for
selection.
SciCal provides this structure.
- pricing, opening times, accessibility, audience expectations (age, club membership, ...), synchronization
(updates?), distinguish information object from event
Politics surrounding data acquisition services. Data ownership issues.
PatH, temporal models
5 years ago
"A survey of temporal formlizations"
James Allen, temporal logics in AI
Survey of formalizations of time
Point, interval, duration, relationship to events
There's no Right Way to do it.
What about points at ends of intervals? Open, closed, ... A. don't ask the question. [[[Use Mereology approach?]]]
Recurrences; interval with pattern of sub-intervals. (Variable hours to match daylight
periods.)
Hierarchy of patterns.
CYC tree; 5th second of 9th minute, of... very general-purpose, maybe too elaborate
for many practical uses.
Time measurements with loosely synchronized clocks?
Future uncertainties? Alternative futures.
[[[Round-trips provide closure]]]
CYC: temporal things vs times.
Directional ambiguity (e.g. 280 N/S/E/W?)
Discontinuity (significance: dinner appointment during a conference)
Causality: do something that stops things happenning.
No universal agreement for resoludion of event identity conflicts. Also, different views of physical
occurrence: are they different events? Do attitudes count?
Is identity the wrong question?
| 
