Re: HTTPS and the Semantic Web

There has always been a tension in RDF arising from the conflation of 
identifiers with locators.  People who want to emphasize a web-like 
character of the "Semantic Web" prefer to use identifiers that also 
function as locators.  I have always thought it better not to do this, 
because a locator can only tell you (possibly changeable) information 
*about* the located thing, rather than being the thing itself (yes, 
there can be cases where they are intended to be the same, but those are 
a small subset of all identifiers).

So we could always have had a property called, say, "hasLocatorString", 
and so we could always have written something like

http://example.com/Concept1 hasLocatorString "http://example.com/Concept1"

In fact, many RDF IRLs tacitly depend on such a statement, but we just 
don't admit it.

With this mechanism, nothing has to change except to specify additional 
"https" locator strings.   As Henry has pointed out, an additional 
subvocabulary around "hasLocatorString" looks like it would be useful as 
well.


On 5/21/2016 9:01 AM, Henry Story wrote:
>
>> On 21 May 2016, at 16:40, Nathan Rixham <nathan@webr3.org
>> <mailto:nathan@webr3.org>> wrote:
>>
>> Why not owl:sameas? Is it technically incorrect?
>>
> <http://xmlns.com/foaf/0.1/knows> owl:sameAs
> <https://xmlns.com/foaf/0.1/knows> .
>
> would be fine, but it would not be clear that one of the URLs is better
> than the other. To do that you need to talk OF the URIs, and not
> the resources, so you need xsd:anyURI. This will do.

Received on Saturday, 21 May 2016 16:33:39 UTC