W3C home > Mailing lists > Public > public-poiwg@w3.org > January 2011

Points of Interest Relationships: Provenance?

From: Christine Perey <cperey@perey.com>
Date: Fri, 21 Jan 2011 13:19:51 +0100
Message-ID: <4D3979E7.2090204@perey.com>
To: "public-poiwg@w3.org" <public-poiwg@w3.org>

The POI WG members are whittle away at definitions of "primitives" which 
were identified by the participants in the FacetoFace meeting as part of 
the Core Principles of the group


Principles 3 and 4 both mention relationships:

point 3 says a POI can have "contained within", "contains" 1-to-many 

point 4 says a POI can have "adjacent-to" relationships

These relationships seem to work well with location POI which are fixed 
in relationship to one another.

Objects do not behave in "fixed" relationships so saying that something 
is contained within something else could break down if the object (e.g., 
the person, or a cup) leaves the building in which it was first found...

Anyway, I don't think that "contained in" or "contained by" or adjacent 
relationships (next to, some distance from) are clearly understood. 
Certainly requires more thought and discussion by people with points of 

The topic of this message is what other types of relationships there 
might be.

A few weeks ago during the WG conf call I asked how the relationships of 
a POI could also have something to do with where the data (metadata, 
attributes of a POI) originated.

And, where the metadata attributed to a POI originated is yet a 
different topic from where the object itself comes/came from, what 
created it (e.g., a manufacturer of a device, provider of a special 
service, a child is the offspring of only one mother and one father, etc).

In circles where I work, we call this provenance.

In the Semantic Web there is a strong concept of provenance. There was a 
W3C Provenance Incubator Group which ended work December 14 2010


This is their final report

And this is their contribution to our understanding of Provenance:

In the social networking space where I've worked extensively advocating 
for open and federated systems, the Social Web Incubator Group asked the 
ProvXG to help us with this topic of provenance:


So...what is the role of provenance in relation to information about 
Points of Interest?

One of the participants of the conference call suggested that the 
provenance of information about a point fixed in geophysical space could 
be said to be the agency which publishes the lat and long. This might 
work for fixed points on the globe but it is definitely not going to 
work for AR uses cases which involve objects that move.

During the conference call on January 19, we struggled with a similar 
problem when discussing the ID primitive.

The question relating to provenance when trying to flesh out what we 
mean by ID is "where is/what is/who is this POI's "parent" ?"

Why should you care?  I can think of many reasons/practical applications.

Under what circumstances will you or your users benefit from knowing the 
provenance of the data which is displayed?

I look forward to your inputs and reactions.


Spime Wrangler

mobile +41 79 436 6869
VoIP +1 (617) 848-8159
Skype Christine_Perey
Received on Friday, 21 January 2011 12:20:23 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 16:48:27 UTC