W3C

- DRAFT -

Wot Discovery Task Force

01 Jul 2015

See also: IRC log

Attendees

Present
DanhLePhuoc, Dave, Danh, Ashok_Helena, Joerg, Hamid
Regrets
Soumya
Chair
Dave
Scribe
dsr

Contents


<DanhLePhuoc> https://filesender.heanet.ie/1.6/?vid=0b60067c-e7bd-c629-bacf-00005974b1bb, the link to the presentation

<scribe> scribenick: dsr

Danh presents Thing Discover on the Graph of Things see link above

Danh: I try to share my demo

Screen shows world map with heat map and locations as overlay

Danh shows how you can browse the data screens using some JavaScript effects

At the bottom there is a live display of data courtesy of a web socket connection with the server.

He returns to the slide presentation

We use HTTP content negotiation as a discovery mechanism, along with a sparql query.

The query results are returned in JSON

We’re using an RDF graph for the data models

There are plenty of ontologies available.

We also have RDF linked data streams

Static knowledge is treated as background knowledge

The sensor metadata describes the stream, with sparql 1.1 specification

We had to extend the sparql query language slightly to be able to aggregate data over a given time period

This also allows you to query the data at a specific time (in the past)

You can define a query that returns a stream for data that satisfies that query

Slide with stats on amount of data, cluster of servers and the software libraries

People ask if you can make it smaller? Yes through smart gateway

We’re working on making it work on Android and Raspberry Pi

The code foot print is in range 4-8 MB.

Memory foot print in range 4 - 64MB

Plenty of room for further improvementsto make it smaller faster and more scalable

Danh finishes

Dave: questions anybody?

Joerg: this looks like a complete solution inspired from sensor networks. We now need to identify some potential building blocks in respect to discovery, any suggestions?

Danh: we only use REST API, complexity is determined by the nature of the query

A few years ago we implemented a small query engine for simple queries

Joerg: in the task force we discussed different notions of discovery

Your presentation is about discovery of certain resource with a particular characteristic.

Dahn: yes, here we focus on data discovery

First you need to discovery the sparql endpoint

The linked data community has ideas for how to query endpoints for their metadata, e.g. their capabilities

We’ve looked at using JSON for queries as an alternative to sparql for restricted classes of queries

Dave: any other questions?

Helena: I want to understand continuuous discovery, is there an anouncement for sensors?

Danh: you first need to discover the query end point, and can then query for all sensors satisfying some constrain.

Helena: the sensors have to register with the system in some way?

Danh: yes

Dave: this is like the Compose project where queries for event streams from a set of sensors is a live query, i.e. reflects the sensors currently registered

Danh: one point - this approach doesn’t have to run on the public Internet.

Technology Landscape for Discovery

Joerg: what are the steps in discovery with links to use cases, and then look at dependencies on work in othe task forces.

You have already listed some mechanisms for discovery, and should now add examples, and the relationship to thing descriptions, APIs and to security, privacy etc.

Dave asks for contributions for next week’s call

Joerg: yes we should invite contributions for the landscape analysis along with links

Dave: I could offer to provide an overview of the Compose system approach to discovery.

Joerg: yes, but you should break it down into the components/building blocks

Summary of Action Items

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.140 (CVS log)
$Date: 2015/07/01 15:50:18 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.140  of Date: 2014-11-06 18:16:30  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Succeeded: s/Dahn/Danh/
Found ScribeNick: dsr
Inferring Scribes: dsr
Present: DanhLePhuoc Dave Danh Ashok_Helena Joerg Hamid
Regrets: Soumya
Got date from IRC log name: 01 Jul 2015
Guessing minutes URL: http://www.w3.org/2015/07/01-wot-di-minutes.html
People with action items: 

WARNING: Input appears to use implicit continuation lines.
You may need the "-implicitContinuations" option.


[End of scribe.perl diagnostic output]