- From: Kazuyuki Ashimura <ashimura@w3.org>
- Date: Sat, 5 Aug 2017 00:18:04 +0900
- To: Public Web of Things IG <public-wot-ig@w3.org>
available at:
https://www.w3.org/2017/08/04-wot-minutes.html
also as text below.
Thanks a lot for taking these minutes, Dave!
Kazuyuki
---
[1]W3C
[1] http://www.w3.org/
- DRAFT -
WoT IG - TF-LD
05 Aug 2017
[2]Agenda
[2] https://www.w3.org/WoT/IG/wiki/IG_Linked_Data_and_Semantic_Processing_WebConf#Agenda
Attendees
Present
Kaz_Ashimrua, Danh_Le_Phuoc, Darko_Anicic, Dave_Raggett,
Michael_Koster, Taki_Kamiya, Victor_Charpenay,
Achille_Zappa
Regrets
Maria_Poveda
Chair
Darko
Scribe
dsr
Contents
* [3]Topics
1. [4]Agenda bashing
2. [5]Shape languages with the focus on Thing Description
data types
* [6]Summary of Action Items
* [7]Summary of Resolutions
__________________________________________________________
<scribe> scribenick: dsr
Agenda bashing
Darko runs through the agenda for today.
[8]Agenda wiki
[8] https://www.w3.org/WoT/IG/wiki/IG_Linked_Data_and_Semantic_Processing_WebConf#Agenda
Dave: we could also talk the work that Maria is doing on the
ontology and my work on JSON Schema as Linked Data.
Darko: I also promised to look at shape languages.
Any additional agenda items for today?
[no]
Shape languages with the focus on Thing Description data types
<DarkoAnicic> type system proposal (currently):
[9]https://github.com/w3c/wot/tree/master/proposals/type-system
[9] https://github.com/w3c/wot/tree/master/proposals/type-system
Darko presents
Darko shows us a side by side comparison of JSON Schema and
schema.org
Darko shows us his experiments (on his pc) with the SHACL and
ShEX RDF shape rules languages
Victor and Darko chat about the possibility of representing
shape rules in JSON
The examples concern validation of data, not the thing
description
ShEx defines its own syntax, but there is a mapping to JSON,
albeit rather verbose
Dave: wonders what we’re trying to achieve in this discussion?
Darko: the idea is to use shape rules to validate the data
Dave: I thought that we were more interested in validating
thing descriptions?
Darko: this is just an exploration of ideas
Dave: I thought we had agreed to create a Linked Data model for
a subset of JSOn Schema, right?
Victor: we’re comparing the constraints available in JSON
Schema with that of SHACL and SheX.
Kaz recaps the discussion and actions from the last face to
face relating to data types, as he is a little confused as to
the current presentation
Kaz: we asked Dave to look at JSON Schema and the web of things
use cases for data types
Darko: I am interested in the role of shape rules in relation
to semantic validation
It seems interesting to compare JSON Schema with shape rule
languages in respect to applying type constraints
(Dave notes that his own work on shape rule languages inspired
by ATNs is being ignored)
<kaz> Kaz: ok. if the purpose for today is rather basic survey
of existing data type/schema languages, that's fine
Darko shows the use of the TopBraid composer tool for use with
SHACL
Dave: my analysis of JSON Schema and WoT use cases:
[10]https://github.com/w3c/wot-thing-description/issues/13
[10] https://github.com/w3c/wot-thing-description/issues/13
Darko: it is interesting to look at the potential for applying
shape rules, e.g. to semantic models
... why is JSON schema so popular, and how does it compare to
shape rules.
<victor> Victor: having a schema that is object-oriented is
important for the scripting API
<victor> because the API itself uses the OO paradigm
Dave: a simple toolkit of techniques for manipulating linked
data makes it easy to apply semantic constraints etc in terms
of operations on sets of nodes and triples
<achille_zappa> [11]https://www.w3.org/TR/shacl-js/
[11] https://www.w3.org/TR/shacl-js/
DanhLePhuoc: It seems there is a lot of confusion here. RDF is
programming language neutral. We can use the right tools for
the task to operate on RDF in the way you want. It is quite
different from object oriented practices
I am also confused as to what outcome you’re seeking in this
discussion
Darko: we want to focus on the data types here
DanhLePhuoc: the scripting API depends on the tasks a
programmer is trying to accomplish
There are lots of libraries for Java, JavaScript etc.
Darko: the only WoT APIs available to us expose objects, and
this makes object oriented techniques relevant
DanhLePhuoc: if developers are using Java, they can use the
Java libraries for linked data etc
I don’t see why Victor want’s to limit the discussion to object
oriented techniques
Victor: the same linked data graph can have multiple JSON-LD
representations
Darko: if you define the JSON-LD context you get a single
translation to linked data
Dave wonders why we aren’t just considering operations on
linked data given that this is the Linked Data and Semantic
Processing Task Force!
mjkoster: I think we need to remember the context for the tasks
He talks about the payload, i.e. the formats used to represent
linked data
The shape rules form one RDF graph that is applied to the data
graph
We need to consider validation of data as a graph, and also the
notation used to serialise it
Darko: couldn’t we use SHACL for thing descriptions?
mjkoster: yes, I see that as a possibility
Dave: let’s put on the agenda for the next call the work I did
on JSON Schema, and use cases for the web of things and the
requirements for linked data.
Darko: sure
taki: we have different worlds to relate (JSON, Linked Data,
shape rules and ontologies)
Darko: any last points for today?
[no]
(Dave: see
[12]https://github.com/w3c/wot-thing-description/issues/13 for
my analysis)
[12] https://github.com/w3c/wot-thing-description/issues/13
[ adjourned ]
Summary of Action Items
Summary of Resolutions
[End of minutes]
__________________________________________________________
Minutes formatted by David Booth's [13]scribe.perl version
1.147 ([14]CVS log)
$Date: 2017/08/04 15:14:11 $
[13] http://dev.w3.org/cvsweb/~checkout~/2002/scribe/scribedoc.htm
[14] http://dev.w3.org/cvsweb/2002/scribe/
Received on Friday, 4 August 2017 15:19:13 UTC