Re: What if an URI also is a URL

On Wed, 06 Jun 2007 16:23:59 -0600, Cédric Mesnage  
<cedric.mesnage@lu.unisi.ch> wrote:

> t is clear a uri must be retrievable both for humans and machines to  
> process. But aren't the solutions of having the server redirecting to a  
> different document based on the content type or using a 303 a bit  
> overkill?

It certainly is overkill if this is the official "directive" in regards to  
how something should be handled.  I don't know enough about the history of  
this particular suggested solution, but I have my doubts that whomever has  
suggested this as a solution did so from a "there should only be one way"  
perspective, and instead a best practices type perspective.  Of course,  
ultimately, regardless of the requesting client, how a particular piece of  
data is accessed should be governed by the existing rules and regulations  
set forth in RFC 2616, and in my own opinion anyone who suggests that  
http:///example.org/foo#bar is a legitimate use for a URI in which, when  
requested, should return the expected resource, in this case 'bar' that is  
member of the 'foo' document clan needs to put down the bottle, take a  
nice long nap or in other forms sober up before they should be allowed  
back at the wheel.

> And will they be followed by everyone, on every blog, wiki or web2.ish  
> webapp of the planet?

What you are alluding to is absolutely spot on!  The obvious answer is:  
No, of course they won't.  Again, as you are alluding to, the only thing  
that can be expected is that when someone makes a request for a particular  
resource on the server, regardless of the syntax of the URI that resource  
should contain nothing more and nothing less than what that particular  
request was expecting.

>
> "All uris should work in browsers", well all urls do,

Nicely stated! :D

> so let us consider a third solution, if the document at   
> "http://www.example.com/mophor" is an HTML document embedding RDF data  
> using RDFa or another way of embedding RDF in HTML about  
> "http://www.example.com/mophor#me", there you get your human and machine  
> readable document for the same thing at the same place. Isn't that  
> wonderful,

And simple too :D

> no need to hack the server to redirect urls, no need for duplication of  
> code to represent data in a way for machines and in another for humans  
> and all used uris are accessible urls.

Amen, brother! :D

>
> Am I on the right track or is there something I missed?

Not only are you on the right track, but the train has already left the  
station and moving full steam ahead.  In fact, that happened 15 some odd  
years ago, and with the momentum that it currently has (meaning the basic  
foundation of what drives the WWW as it exits today (AKA DNS, (U,I)RI,  
(X,(X)HT)ML and HTTP)), any attempt to redefine how data should be  
interpreted by the client when using these protocols/formats is a horrific  
disaster waiting to happen.

-- 
/M:D

M. David Peterson
http://mdavid.name | http://www.oreillynet.com/pub/au/2354 |  
http://dev.aol.com/blog/3155

Received on Wednesday, 6 June 2007 23:43:22 UTC