W3C home > Mailing lists > Public > public-prov-wg@w3.org > August 2011

RE: PROV-ISSUE-72 (DGarijo): Uses should be renamed as used [Formal Model]

From: Ralph Hodgson <rhodgson@topquadrant.com>
Date: Wed, 24 Aug 2011 11:04:43 -0700
To: "'Simon Miles'" <simon.miles@kcl.ac.uk>, "'Provenance Working Group WG'" <public-prov-wg@w3.org>
Message-ID: <015701cc6288$4d83cf60$e88b6e20$@com>
Replies in-line.

 

Ralph Hodgson,  <http://twitter.com/ralphtq> @ralphtq

Mobile Phone: +1 781-789-1664

CTO,  <http://www.topquadrant.com/> TopQuadrant,
<http://twitter.com/topquadrant> @TopQuadrant

Blog:  <http://topquadrantblog.blogspot.com/> Voyages of the Semantic
Enterprise

 

From: public-prov-wg-request@w3.org [mailto:public-prov-wg-request@w3.org]
On Behalf Of Simon Miles
Sent: Wednesday, August 24, 2011 2:20 AM
To: Provenance Working Group WG
Subject: Re: PROV-ISSUE-72 (DGarijo): Uses should be renamed as used [Formal
Model]

 

With regard to tenses, I note that the current relations are not uniformly
in the same tense anyway.

 

The problem is that "derived" and "generated" are different kinds of word -
"derived" can be a contemporary property making reference to the past, while
"generated" is always a past tense action.

 

"X is derived from Y" can be read as saying, "it is a property of X that it
is, at this moment, derived from Y", e.g. "my name is derived from ancient
Norse", and this has the meaning intended by our model.



[RH]  is it not the case that 'derivedFrom' (at  least for this ontology)
implies a past event? I suppose there is a subtle distinction between
'isDerivedFrom' (when used to explain how an algorthim works, for example)
and 'wasDerivedFrom' (when justifying some belief), and maybe a requirement
in some ontology, perhaps not a provenance ontology, to say
'willBeDerivedFrom'?

 

"X is generated by Y" only has one meaning in English which is "whenever
instances of class X occur they were, are and will be generated by instances
of class Y", e.g. "smoke is generated by fire". This is not the meaning
intended by our model and is not about provenance. Similarly for
"controlled".

 

It may be a small thing and we could say that the model should be read
without pedantic observance of English grammar, but it is just something
else which makes the model harder to understand and apply correctly. I
personally see least ambiguity in uniformly past tense: "wasGeneratedBy",
"used", "wasDerivedFrom"... For users who don't yet have a firm grasp of
what "provenance" is, past tense can greatly help convey the intuition.

[RH] I am beginning to agree to using 'was' for being clear on past tense.
If provenance has to consider present and future assertions then we add more
properties? But I don't know the requirements we might have for dealing with
present and future in a provenance ontology. I subscribe to the Antoine de
Saint-Exupery principle of  "A designer knows he has achieved perfection not
when there is nothing left to add, but when there is nothing left to take
away." Therefore, hope we do not have to add the other tenses.

Thanks,

Simon

 

On 23 August 2011 21:12, Deborah L. McGuinness <dlm@cs.rpi.edu> wrote:

On 8/23/2011 3:35 PM, Satya Sahoo wrote: 

Hi Ralph, 

My comments are inline:

 

> would recommend dropping the 'has' from 'hasParticipant' (silent 'has'
rule on noun predicates)

i strongly support naming conventions that help people know at a glance if a
term is a property/role or a class/concept.
Thus I rarely support a "silent has" rule.

I have found over many years of teaching and consulting that it is too easy
for people to misuse classes as properties as vice versa but when naming
makes it clear which is which, they make way less errors and reuse is
increased.





, or, optionally, finding an appropriate gerund phrase with a noun suffix,
perhaps 'Party' in this case. The predicate would then be
'participatingParty'.

 

For the participation property, a verb is required since - (a) gerund phrase
is continuous (participating) while other properties you suggested are
perfective (generated, controlled), and (b) in a triple construct, it will
read as "(some) process execution" -> participatingParty -> "(some) entity",
which does not convey the intended meaning of the property.

 

I think we should keep hadParticipant, but use other property names you have
suggested. 

 

> Specification of tense can be done either with reification or, in the case
of, RDF-based models, use of rdf:Statement.

I did not understand this, can you please give an example?

 

> Also advocate that all predicates begin with a lowercase letter.

We already do that as mentioned in the formal model document.

i agree with the lower case .
some conventions use lower case for properties and upper case for classes.




 

Thanks.

 

Best,

Satya

 

On Mon, Aug 22, 2011 at 5:24 PM, Ralph Hodgson <rhodgson@topquadrant.com>
wrote:

Hi Luc,

 

I would recommend dropping the 'has' from 'hasParticipant' (silent 'has'
rule on noun predicates), or, optionally, finding an appropriate gerund
phrase with a noun suffix, perhaps 'Party' in this case. The predicate would
then be 'participatingParty'. So in answer to your question, the predicate
names become tense-less verb or gerund phrases.  Specification of tense can
be done either with reification or, in the case of, RDF-based models, use of
rdf:Statement.

 

Below is a diagram generated by TopBraidComposer of, what I believe to be a
current Provenance Ontology model )from
http://dvcs.w3.org/hg/prov/file/738e9b4d8520/ontology/ProvenanceOntology.owl
):

 



 

The Turtle file is attached (I need to consult others on where to put such
files on the WG site). I transformed the OWL original using SPIN rules so I
can repeat this process in the future.

 

Ralph Hodgson,  <http://twitter.com/ralphtq> @ralphtq

Mobile Phone: +1 781-789-1664 <tel:%2B1%20781-789-1664> 

CTO,  <http://www.topquadrant.com/> TopQuadrant,
<http://twitter.com/topquadrant> @TopQuadrant

Blog:  <http://topquadrantblog.blogspot.com/> Voyages of the Semantic
Enterprise

 

From: Luc Moreau [mailto:L.Moreau@ecs.soton.ac.uk] 
Sent: Monday, August 22, 2011 1:57 PM
To: rhodgson@topquadrant.com
Cc: public-prov-wg@w3.org


Subject: Re: PROV-ISSUE-72 (DGarijo): Uses should be renamed as used [Formal
Model]

 

Hi Ralph,
Don't you end up with a mix of relation names:
 uses, derivedFrom and hasParticipant ?

My preference is for uniformity. If we go for a verbal form for some, we
should go for all;
otherwise for none.

Luc

On 22/08/11 20:09, Ralph Hodgson wrote: 

Propose that 'is' be dropped from all predicates that have a suffix. Two
reasons - parsimony and connotations (in fact denotations) about tense are
removed.

 

            Example: 'isDerivedFrom' becomes  'derivedFrom'.

 

Also advocate that all predicates begin with a lowercase letter.

 

Ralph Hodgson,  <http://twitter.com/ralphtq> @ralphtq

Mobile Phone: +1 781-789-1664 <tel:%2B1%20781-789-1664> 

CTO,  <http://www.topquadrant.com/> TopQuadrant,
<http://twitter.com/topquadrant> @TopQuadrant

Blog:  <http://topquadrantblog.blogspot.com/> Voyages of the Semantic
Enterprise

Next Intro Class: Introduction to Semantic Web Technologies: What they are
and how to use them - <http://www.topquadrant.com/training/intro.html>
Sept. 12 - 15, 2011, Washington DC area , 

Next Advanced Class: Advanced Product Training, Sept. 26 - 29 , 2011,
Washington, DC area <http://www.topquadrant.com/training/advanced.html> 

 

From: public-prov-wg-request@w3.org [mailto:public-prov-wg-request@w3.org]
On Behalf Of Luc Moreau
Sent: Monday, August 22, 2011 2:53 AM
To: public-prov-wg@w3.org
Subject: Re: PROV-ISSUE-72 (DGarijo): Uses should be renamed as used [Formal
Model]

 

Hi Daniel and Simon,

In the conceptual model (up to section 5), all relations use present tense,
uses, isDerivedFrom, isComplementOf, isGeneratedBy, ...

I agree with Daniel, we need consistency in the conceptual model and formal
model.

I agree with Simon, it's nice to point back into the past.

Luc

PS. I took as an action to poll about the use of tense.

On 08/11/2011 04:46 PM, Simon Miles wrote: 

+1 for "used", for the reasons you give.
 
-1 for "isUsedBy". All other relations in the model link effect/later
occurrence to cause/earlier occurrence, so "isUsedBy" would be
inconsistent. This means that provenance graphs would not consistently
point back into the past of the thing we're seeing the provenance of,
and would be confusing to iteratively navigate.
 
Thanks,
Simon
 
On 11 August 2011 16:38, Provenance Working Group Issue Tracker
 <mailto:sysbot+tracker@w3.org> <sysbot+tracker@w3.org> wrote:
  

 
PROV-ISSUE-72 (DGarijo): Uses should be renamed as used [Formal Model]
 
http://www.w3.org/2011/prov/track/issues/72
 
Raised by: Daniel Garijo
On product: Formal Model
 
Since all the other properties are using the past tense, we should be
consistens and rename uses to "used" or "isUsedBy" flipping the domain and
range
 
 
 
 
______________________________________________________________________
This email has been scanned by the MessageLabs Email Security System.
For more information please visit http://www.messagelabs.com/email
______________________________________________________________________
 
    

 
 
 
  





-- 
Professor Luc Moreau               
Electronics and Computer Science   tel:   +44 23 8059 4487
<tel:%2B44%2023%208059%204487>          
University of Southampton          fax:   +44 23 8059 2865
<tel:%2B44%2023%208059%202865>          
Southampton SO17 1BJ               email: l.moreau@ecs.soton.ac.uk  
United Kingdom                     http://www.ecs.soton.ac.uk/~lavm
<http://www.ecs.soton.ac.uk/%7Elavm> 

 

 





 

-- 
Dr Simon Miles
Lecturer, Department of Informatics
Kings College London, WC2R 2LS, UK
+44 (0)20 7848 1166


image001.png
(image/png attachment: image001.png)

Received on Wednesday, 24 August 2011 18:05:04 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 16:51:00 UTC