On 12/14/2013 2:42 AM, davep wrote:
> On 13/12/13 18:26, Liam R E Quin wrote:
>
>>
>> I can't easily say, using RDF,
>> http://www.fromoldbooks.org/Aubrey-HistoryOfEngland-Vol2/pages/438-detail-Portrait-of-King-Henry-VIII/#fg=%237929ce_bg=none
>>
>> is an image of a wood-engraving depicting King Henry VIII. of England
>> which has been coloured purple.
>>
>> To do that I need URIs for King Henry and for England and for purple, as
>> opposed to URIs for Web pages about those things.
>>
>> The HTTP Range discussion I mentioned was an attempt to say that a # on
>> the end of a URI meant you were using the URI as a surrogate for a
>> person (or was it the absence of a #? I forget).
>>
>> Hope this helps.
>>
>> Liam
>>
>
> Can't you? Surely at some level, the content of the <a/> element COULD
> do this?
>
> <a href='Liams web'>an image of a wood-engraving depicting King Henry
> VIII. of England which has been coloured purple.</a>
Yes I think that's my position exactly. Why are we distinguishing
between humans and machines, and more - privileging the machines?
Machines are doing more and more with our language: I think that is the
direction to push: helping the machines to understand us, rather than
attempting to understand what it is the machines want from us.
To me the introduction of an additional attribute (rel, rev, whatever)
beyond the text just pushes the same problem off into a different
syntactic construct - what's the point? It's a little bit like thinking
that namespace uris are going to be more unique than namespace
prefixes. It just adds complex syntax without doing anything to solve
the essential problem.
-Mike
