W3C home > Mailing lists > Public > public-vocabs@w3.org > September 2014

RE: New Proposal - Measurement (in support of describing sports statistics)

From: <Simon.Cox@csiro.au>
Date: Mon, 8 Sep 2014 06:29:49 +0000
To: <jasjoh@microsoft.com>, <martin.hepp@ebusiness-unibw.org>, <tom.grahame@bbc.co.uk>
CC: <public-vocabs@w3.org>
Message-ID: <2A7346E8D9F62D4CA8D78387173A054A5FFFB93E@exmbx04-cdc.nexus.csiro.au>
> a Measurement appears to be primarily a means of tying an Event to a Quantity 

That's essentially how it is modelled in O&M: a Measurement is n-ary relationship between the feature-of-interest, a property of the feature, a (estimate of the) value of the property, and a procedure used to determine (a.k.a. estimate) it. 
This is the value-assignment viewpoint, which is primarily important if you care about the procedure-application-event because you care about uncertainties in the value of the result. 

There are other viewpoints, in particular 
- feature viewpoint - the thing in the world with all its properties and their values, without caring how they were estimated;
- procedure or sensor viewpoint - the sensor/platform/procedure and all the results it generated - which tends to be preferred by people who have a rare and expensive one (like a ship or a satellite or a supercomputer with a complex algorithm), so merely naming it conveys a lot of information to the cognoscenti. 

Simon 

-----Original Message-----
From: Jason Johnson (BING) [mailto:jasjoh@microsoft.com] 
Sent: Friday, 5 September 2014 11:03 AM
To: Jason Johnson (BING); Cox, Simon (L&W, Highett); martin.hepp@ebusiness-unibw.org; tom.grahame@bbc.co.uk
Cc: public-vocabs@w3.org
Subject: RE: New Proposal - Measurement (in support of describing sports statistics)

Further .... 

I am struggling with modeling measurements because of a fundamental issue I have with the relationship between events, actions, and results and my underlying desire powering all of this work, which is the desire to describe sports statistics.  I think I've had a minor break through and since my brain has had about enough for the day, I'm going to share my latest thoughts as is.  

============================================

Event is a well-defined / named duration in time Event can encompass the performance of one or more Actions; Action Instances Event can represent a single instance of an Action; a single performance of an Action Event sometimes have spatial properties as well by way of Actions 

The Result of an Event is the sum of all the Results of the Actions it encompasses

The Result of an Action is a change in one or more objects and their respective state, including creation and destruction of entire objects.  The Result of an Event is a set of these Action initiated state changes.

Statistics are the meta data resulting from collection, organization, analysis, and interpretation of a set of data.  In the case of Event Statistics, the counts and calculations are usually based on the set of state changes that take place during an event.

Within the scope of Major League Baseball, statistics are related to a baseball player by way of the actions they perform and the events in which they perform those actions.  Baseball statistics are always associated with an event, even when that event is abstract or broad as their 'career' or 'lifetime'.  Similarly, they are related to baseball teams by way of the collective actions of the players of that team.  Perhaps a stretch, but you could inversely state that they are related to parts of a body as well, by way of the player being a collection of body parts.  Think 'batting left handed' vs 'batting right handed'.

===============================================

Tying this back to measurements but maintaining the scope of sports (Event) statistics, a Measurement appears to be primarily a means of tying an Event to a Quantity.  In other words, the action of measuring the set of state changes in an event resulted in a establishing a quantity which was an event statistic, which in turn was used as a property to describe the performance of an athlete or sports team.

Turning off that part of my brain until tomorrow ...



-----Original Message-----
From: Jason Johnson (BING) [mailto:jasjoh@microsoft.com]
Sent: Thursday, September 4, 2014 5:43 PM
To: Simon.Cox@csiro.au; martin.hepp@ebusiness-unibw.org; tom.grahame@bbc.co.uk
Cc: public-vocabs@w3.org
Subject: RE: New Proposal - Measurement (in support of describing sports statistics)

My latest thinking:  

Measurement, in its noun form, consists of a quantitative value which was assigned to a quantity representing a property of the direct object of the Measurement action.  In addition, Measurement, in noun form, consists of a bunch of other meta data associated with the performed act of Measurement, such as when the measurement took place, how long it took, what instruments were used, what conditions were in place at the time of the measurement, etc.

A Quantitative Value (Quantity Value / Quantity) is simply a number and a unit of measurement.  That's it. 

If you want to capture just a number and unit, use a Quantitative Value.
If you want to capture the semantics of how that Quantitative Value was established, use a Measurement.



-----Original Message-----
From: Simon.Cox@csiro.au [mailto:Simon.Cox@csiro.au]
Sent: Thursday, September 4, 2014 5:16 PM
To: martin.hepp@ebusiness-unibw.org; tom.grahame@bbc.co.uk
Cc: public-vocabs@w3.org
Subject: RE: New Proposal - Measurement (in support of describing sports statistics)

> there is no relevant difference between a quantitative value and a 
> measurement

As defined in VIM and O&M, and reflected in the current schema.org proposal, these are quite different things. 
'Measurement' is an action or event, the _result_of_which_ is a quantitative value. 

Simon Cox

-----Original Message-----
From: martin.hepp@ebusiness-unibw.org [mailto:martin.hepp@ebusiness-unibw.org]
Sent: Thursday, 4 September 2014 11:10 PM
To: Tom Grahame
Cc: W3C Web Schemas Task Force
Subject: Re: New Proposal - Measurement (in support of describing sports statistics)

Conceptually, I think there is no relevant difference between a quantitative value and a measurement, since all what we process in computer systems are abstractions of reality, so any qualitative value represented in any computer system is essentially tied to a measurement activitity (except for values generated randomly or algorithmically, but with a wider notion of what "measurement" means, that could well be covered, too).

Martin



On 04 Sep 2014, at 14:59, Tom Grahame <tom.grahame@bbc.co.uk> wrote:

> 
> On 04/09/2014 10:51, "Markus Lanthaler" <markus.lanthaler@gmx.net> wrote:
> 
>> - Is quantityValue really required? Can't a Measurement be a 
>> QuantitativeValue at the same time?
> 
> As I interpret the examples, the Measurement is the activity (a form 
> of
> observation) and the QuantitativeValue an outcome of the activity, so 
> they are different things.
> 
> Iım not sure what is correct but the problem may stem from this part 
> of the proposal, in the Terminology section, where I think Measurement 
> may need to settle one way or the other:
> "The term Measurement may refer to act of measuring (its verb form) or 
> the result of that act (its noun form).²
> 
> Tom
> 
> 
Received on Monday, 8 September 2014 06:30:31 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:29:44 UTC