Re: Site metadata; my preference

Patrick.Stickler@nokia.com wrote:
> 
> ...
> 
> Oh, come now. Were the authors of that statement thinking "GET vs. any
> possible arbitrary verb that might every be added to HTTP" or "GET vs. POST"?!

Actually it was probably "GET vs. any possible arbitrary verb that might 
ever be added to SOAP." The point is that the Web has a _single_ 
namespace. Every URI identifies one resource. If you put an MGET there, 
then you've got two resources at the same URI: "X" and "metadata about 
X". What if I want to refer to "metadata about X" as its own first-class 
resource, do I need a whole new URI scheme?

> I'm going to be hard to convince that it was the former.
> 
> And this argument is not even applicable against MGET because with
> MGET documents *are* identified by URI!
> 
>  HTTP GET  {URI} -> resource
>  HTTP MGET {URI} -> description

Look, if I want to say that {URI} is wrong, then I can express that

{URI} <full:of> shit

Now what if I want to say that the URI's metadata is wrong? If I don't 
have a URI for that metadata, I can't do it. What if I want to say that 
the document is authored by one person but the metadata is authored by 
another?

You've said that "sometimes" the metdadata could have its own URI. This 
implies that the person owning the metadata can choose whether other 
people can make statements about it. But that's not the way the Web 
works. If the thing is on the Web, people should be able to make 
statements about it, whether the owner wants them to or not.

  Paul Prescod

Received on Tuesday, 18 February 2003 20:42:55 UTC