WP6 face-to-face meeting in Stuttgart on 2008-10-14
Participants:
- Cefriel: Emanuele, Irene
- Saltlux: Kono, Tony, Stanley
- VUA: Zhisheng
Discussion
- M9 - D6.2 - Templates of periodic report on data and performances
- Stefano Bertolo asked for measuring the progress of the work in the project wrt parameters and metrics over the use case
- from the discussion with WP5, we can think about QoS parameters and things like those
- this will also be used for M12 - D6.4 - 1st periodic report on data and performances
- M12 - D6.3 - Urban Computing environment specification
chicken&egg problem with technical WPs, in particular with WP5
Dieter proposal in TMB meeting: just a mock-up & paper for M12
change the name to "First Urban Computing Service Mockup" --> demonstrator, data and paper
in this case, we must have a plan for the mockup within M8 --> (1) to use Saltlux UCity with Larkc; (2) to build www.linkedurbandata.org exposing as SPARQL endpoint the public data we collected; (3) possibly, to overlap the two things at least partially
- we can also use this mockup as a WP4 plugin test (as suggested by Zhisheng), as a C-SPARQL test (for the Cefriel-Polimi group working on it), etc.
- TASKS: (1) what are those (to-be-linked) urban data? (2) what is the storyboard behind the mockup? starting from the current storyboards
- Current storyboard steps and requirements for the mockup (cf. Emanuele presentation)
- just a form to enter data
- some various constraints:
- we don't have data ready from Italian railways, subway and car traffic
- we do not use temporal reasoning (no ontology of time)
- we have to check a place where we have timetables of train and subway, road network (street topology) and traffic
but without information like commuters traffic, i.e. time needed to follow a link on the path --> maybe Siemens can compute the "weight" of the links for different periods of the day? only if loosely coupled
- multiple-transportation travel planning (even if something is made by us and something invoked from external services),
- traveling costs, user preferences (e.g., first class rather than second class on trains, cheapest, fastest, ...);
--> priority of choices: (1) check the data, (2) everything without travel costs and user preferences, (3) if we succeed, also costs and preferences
- ok
the system stores the travel plans and alert if something changes and makes the chosen option no longer valid --> it is a trigger fired by some change in the assumptions; no problem for the mockup (it is a user preference)
is it possible to get "real-time" updates of the data to be alerted? we need a way to "feed" the system with e.g. accident data --> ok, but with faked events (not real, but realistic and inserted on purpose)
no, this is very hard --> simplify: a delay invalidates the plan
- ok, with (optional) explanation of why the plan is no longer valid (with fake SMS to alert and without the parking ticket thing)
- not the SMS, but the user checks the alternatives on the web interface (e.g. on his smart phone)
- maybe, but only if Siemens wants to create a separate prediction demonstrator with the weather as well
- ok
- it is not necessary to find the cheapest gas station
- no season/holiday/weather conditions, but car accident and road re-planning (one plan is mandatory, multiple alternative plans are optional)
- ok
no metro, but Carlo arrives in time
