Re: What is a URL? And What is a URI

On 11/12/10 8:40 AM, Patrick Durusau wrote:
> Kingsley,
>
> On Fri, 2010-11-12 at 07:58 -0500, Kingsley Idehen wrote:
>> On 11/12/10 5:59 AM, Patrick Durusau wrote:
> <snip>
>
>>>
>>>
>> Patrick / Dave,
>>
>> I am hoping as the responses come in we might pick up something. There
>> is certainly some confusion out there.
>>
>> Note my comments yesterday re. URIs and Referents. I believe this
>> association to be 1:1, but others may not necessarily see it so.
>>
> Isn't it that "...others may not necessarily see it so." that lies at
> the heart of semantic ambiguity?

Yes!

We are perpetuating ambiguity by conflating realms, ultimately. The Web 
of URLs != Web of URIs. They are mutually inclusive (symbiotic).

> Semantic ambiguity isn't going to go away. It is part and parcel of the
> very act of communication.

This is why Context is King.

You can use Context to reduce ambiguity.

A good Comedian is a great Context flipper, for instance.

Ambiguity exists in the real-world too, we use Context to disambiguate 
every second of our lives.

> It is true that is very limited circumstances with very few semantics,
> such as TCP/IP, that is it possible to establish reliable communications
> across multiple recipients. (Or it might be more correct to say
> semantics of concern to such a small community that agreement is
> possible. I will have to pull Stevens off the shelf to see.)
>
> As the amount of semantics increases (or the size of the community), so
> does the potential for and therefore the amount of semantic ambiguity.
> (I am sure someone has published that as some ratio but I don't recall
> the reference.)

So if a community believes in self-describing data, where the data is 
the conveyor of context, why shouldn't it be able express such believes 
in its own best practice options? Basically, we can solve ambiguity in 
the context of Linked Data oriented applications. Of course, that 
doesn't apply to applications that don't grok Linked Data or buy into 
the semantic fidelity expressed by the content of a structured data 
bearing (carrying) resource e.g. one based on EAV model + HTTP URI based 
Names.

> Witness the lack of uniform semantics in the linked data community over
> something as common as sameAs. As the linked data community expands, so
> are the number of interpretations of sameAs.

IMHO. There are a number of people that postulate about owl:sameAs on an 
"its my way of the highway" basis. In my experience, said position 
reflect partial understanding of Linked Data and the ability to 
inference conditionally. Yes, you can inference conditionally if 
inference only occurs when a collection of rules are actually applied, 
and the platform in question offers this kind of deftness. There isn't a 
rule that says all inference must be forward-chained, unfortunately, 
many of the people that gripe about owl:sameAs (good, bad, and damn 
right ugly) do so assuming:

1. Linked Data and Inference are inextricably bound
2. Inference must be forward-chained as opposed to having a 
backward-chained option
3. Inference can't be optional.

Anyway, Harry, Pat et al. have a great paper [1] on owl:sameAs (you may 
have read this already). In addition there's a nice Similarity Ontology 
[2] that does the opposite of endless postulation, it just adds 
granularity by offering a collection of terms people may or may not opt 
to use.
> Why can't we fashion solutions for how we are rather than wishing for
> solution for how we aren't?

I think Ian's option addresses that problem.
> Hope you are looking forward to a great weekend!

Yes! Hope the same for you :-)

Links:

1. http://iswc2010.semanticweb.org/pdf/261.pdf -- Harry Halpin, Patrick 
J. Hayes, James P. McCusker, Deborah L. McGuinness, and Henry S. 
Thompson re. owl:sameAs field use
2. http://bit.ly/aBaPHt -- Similarity Ontology Overview

> Patrick
>
>
>
-- 

Regards,

Kingsley Idehen 
President&  CEO
OpenLink Software
Web: http://www.openlinksw.com
Weblog: http://www.openlinksw.com/blog/~kidehen
Twitter/Identi.ca: kidehen

Received on Friday, 12 November 2010 15:13:12 UTC