Re: [metaDataInURI-31]: Initial draft finding for public review/comment.

In a message dated 7/9/2003 8:26:56 AM US Mountain Standard Time, 
paul@prescod.net writes:

> First, additional metadata in a URI makes it
> >easier to keep the URI "cool" (as in 
> >http://www.w3.org/Provider/Style/URI.html) by
> >adding classification metadata to the identifier (as with the W3C URLs 
> >in your finding).
> 
> I disagree. The more semantic information is in a URI, and the more 
> clients depend upon that semantic information, the harder it is for you 
> to change the URI when the metadata changes. e.g.
> 
> http://foo/document_about_URLs_by_Paul.html
> 
> What happens when Mike takes over editing? Or the document changes to be 
> about URLs, not just URIs? Or XML, not HTML?
> 

While including semantic information in a poorly designed scheme
could cause URIs to be brittle, much of that brittleness depends on 
the type of metadata chosen to include in the URI.  Temporary or 
volatile pieces of information are not good candidates.  Topics are 
good candidates (as in your example -- URI or URL).  In the case of 
the topic changing from URI to URL, the old document should be 
archived in its original form.


> >Second, additional metadata in a URI enables a higher-level
> >of efficient processing on resources by applications that *just* want
> >to process URIs.  Opaque URIs would eliminate that increasing possibility.
> 
> That's true, but you're trying to use the URI to do two different jobs 
> at once (identify and carry metadata). You could instead invent a new 
> syntax for URIs + metadata. The URI part would be the true identifier 
> and would be very persistent. The metadata part would change as the 
> resource evolved.
> 

I agree that I am trying to use the URI for two purposes; however,  
identifiers often perform both of those purposes in the real world.  If I 
call someone "Dr. Joseph Jones".  I have added metadata into the 
identifier because it is very important semantic information to convey
rapidly.  The issue seems to be around restricting the type of metadata
in a URI to two areas: critical semantic information that reflects the nature
of the thing's identity and non-volatile information.

 - Mike
---------------------------------------------------
Michael C. Daconta
Chief Scientist, APG, McDonald Bradley, Inc.
www.daconta.net