Brass Tacks

I've been following the range-14 discussion, which is why I joined this. 
Hopefully I've followed it OK. I think the principle of trying to adapt 
the rules to the cases "in the wild" is a smart move (like making paths 
where there's already a mud track). RDF is really hard for people* to 
understand, anything that makes it easier to get right would be a huge win.

I'm trying to think how to dumb down this explanation so I can think 
about it clearly; Is the following an accurate summary?

"In this suggestion, when you resolve a URL and get "200 OK" you get a 
document of some kind. This document is *Located* by the URI, but not 
necessarily *Identified* by the URI. HTTP Link headers can be used to 
clarify the exact relationship between the document returned, the URI 
requested, and other documents relating to the URI or document."

So this seems to be moving the (required) ambiguity of "303 See Other" 
directly into "200 OK", which makes sense to me, what with content 
negotiation meaning different people might get different documents anyway.

This means that when you resolve a URL you get some stuff which might 
have some association with it as a URI, if it has a URI. Currently URL 
is a sub-set of URI. If I've understood correctly, this would make it an 
intersecting set; Not all URIs would locate something, not all URLs 
would identify something.

Far from worrying me, this vagueness seems to better specify how the web 
is used; you can be strict and specific if you want, but if you're not 
then it'll still work but people shouldn't make assumptions.

The "All URLs are URIs" has been a fundamental law of nature for most of 
my adult life. But it could be this is the sacred cow causing all the 
bother.


* I know saying "RDF is hard" is liable to make a bunch of people tell 
me I'm wrong, but it's certainly harder than it needs to be, and hard 
enough not to provide a return-on-investment for on-the-ground web 
developers to learn. My personal success criteria is the year we start 
seriously having to address the problem of "semantic spammers". Spammers 
and frauders etc. don't do anything unless there's an ROI on the effort!



-- 
Christopher Gutteridge -- http://id.ecs.soton.ac.uk/person/1248

Graphite PHP RDF Library v1.5 released! http://graphite.ecs.soton.ac.uk/
University of Southampton Open Data Service: http://data.southampton.ac.uk/
You should read the ECS Web Team blog: http://blogs.ecs.soton.ac.uk/webteam/

Received on Tuesday, 27 March 2012 08:37:27 UTC