Re: URI vs Relative URI

In a message to atom-syntax@imc.org Larry Masinter wrote:

> Some people have been concerned about the community
> confusion on the terms "URI" and "Relative URI".
> 
> The question is whether anyone thinks the document
> 
> http://www.ietf.org/internet-drafts/draft-fielding-uri-rfc2396bis-06.txt
> 
> is not clear on the definitions of these terms and
> the usage of the protocol elements.
> 
> If you have comments about that particular document,
> please reply to uri@w3.org.

It is not clear.

The statement "Relative URIs are not URIs" will sound like complete 
nonsense to most implementors. I didn't fully understand this 
distinction until recently, when an expert on the matter took the time 
to explain it to me.

Most people think an absolute URI is anything with a scheme and host (a 
fragment is part of an absolute URI), and, together with Relative URIs, 
they make up the set of URIs.

I noticed that the spec has clarified "common misunderstandings" in 
other areas. I encourage the inclusion of similar language for this issue.

Robert Sayre