W3C home > Mailing lists > Public > public-semweb-lifesci@w3.org > June 2007

Re: Evidence

From: Alan Ruttenberg <alanruttenberg@gmail.com>
Date: Wed, 20 Jun 2007 14:04:11 -0400
Message-Id: <89BD53F6-C4DE-43CF-A6E0-48BDF81C2730@gmail.com>
Cc: "Kashyap, Vipul" <VKASHYAP1@PARTNERS.ORG>, <public-semweb-lifesci@w3.org>, "Pat Hayes" <phayes@ihmc.us>
To: "Mark Montgomery" <markm@kyield.com>

On Jun 20, 2007, at 12:06 PM, Mark Montgomery wrote:

> Alan,

Hi Mark,

> Rather than using the public community for just debate, perhaps it  
> would be best used as an educational tool.

Ouch :)  (though I do think that debate is a useful tool in  
education, and a survey of my posts would demonstrate that I've used  
the list for a number of purposes)

> For example, what is your definition of a construct within the  
> context of knowledge systems design in life sciences?

First, let me note that I did not originate the use of the term in  
this discussion

Constructs occur at many levels. In this conversation the issue that  
is being raised is the utility of distinction between continuant and  
occurrent, or roughly, thing versus process (stuff that happens to  
things), as the distinction is made in BFO and many other ontologies.

The counterproposals are not entirely clear to me, but in the case of  
Patrick I believed that the alternative is to abandon this  
distinction and consider everything a process, which may work equally  
well - I don't have enough experience to know.

In other areas I would use construct in different ways: WRT OWL  
versus other KR/Logic languages I might call a construct the method  
of writing an existential as a someValuesFrom restriction.

For the discussion of evidence I would call the method I proposed -  
that in the case of the gene ontology annotation, the evidence is  
about the existence class, and that the evidence/provenance is  
associated with that class - a construct. Or a common agreement about  
how we will say the information in an entrez gene record is about, or  
mechanism for representing process that takes advantage of open world  
semantics to provide room, without revising what is originally said,  
to add more detailed process information if we happen to come to know  

I'm most interested in the last class of constructs, though I try  
understand and work with the rest in order to be able to make systems  
that do something.

> I have found that occasionally revisiting definitions to include  
> refinement is essential for advancement of a practice, which has  
> indeed been a sticky point in knowledge systems often enough. And  
> even when agreement can be reached, definitions tend to be fluid,  
> particularly in fast evolving areas of science.

No disagreement.

> Also, for those who use generic email addresses without links to  
> web sites, it would be very useful to occasionally inform folks on  
> our backgrounds and relationships, like a link to a web page and/or  
> bio for example.

OK, regards,

Alan Ruttenberg
Principal Scientist, Neurocommons,
Science Commons

Some things about me and projects I work on:


> .02
> Mark Montgomery
> CEO, Kyield
> http://www.kyield.com
> Managing Partner
> Initium Venture Capital
> http://www.initiumcapital.com
> ----- Original Message ----- From: "Alan Ruttenberg"  
> <alanruttenberg@gmail.com>
> To: "Kashyap, Vipul" <VKASHYAP1@PARTNERS.ORG>
> Cc: <public-semweb-lifesci@w3.org>; "Pat Hayes" <phayes@ihmc.us>
> Sent: Wednesday, June 20, 2007 8:53 AM
> Subject: Re: Evidence
> On Jun 20, 2007, at 10:08 AM, Kashyap, Vipul wrote:
>> VK> I think Pat raises a very valid point. I do sense that folks   
>> are treating BFO as “true” and sometimes probably trying to shoe-  
>> horn their requirements into it. The scientific method requires   
>> that the various constructs and distinctions of BFO (and for that   
>> matter others such as DOLCE, OpenCyc, etc.) be evaluated from the   
>> use perspective and see if they bring any value in the context of  
>> a  real application probably healthcare and life sciences.
> Usually I think of the scientific method as trying to determine
> truth, not utility, though as you know I'm a big one for utility.
> Note that my general support for BFO has been based on its utility in
> collaboratively building ontologies for combining knowledge,
> particularly OBI and in the HCLS demo. I'm quite interested in anyone
> else's work that might be used to be able to evaluate alternatives,
> but I plan to invest my limited time in continuing to use and improve
> BFO until it breaks in a way that can't be fixed.
>> BTW, I do not intend to evaluate whether these constructs are  
>> ontologically sound, etc. but the question we need to answer as a   
>> group is: Are these constructs useful?
> I don't know what "ontologically sound" means. I would offer that a
> "best practice" would be be to make sure that part of our "acceptance
> tests" for agreeing that something is useful is that many of us
> understand what is meant by a construct.
> -Alan
Received on Wednesday, 20 June 2007 18:04:20 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 17:20:27 UTC