Re: Excess URI schemes considered harmful

On Tue, Sep 25, 2001 at 04:58:44PM -0700, Roy T. Fielding wrote:
> > At the end of the day, it would be great if this community could agree on 
> > *one* mapping between the two formats.  I'd get laughed out of the room if 
> > I suggested as a work item for our developer team to map the n-to-n 
> > possible ways a media type can be expressed as a URI as something we should 
> > implement in any of our products.
> Me too.  Personally, I think it is absurd to require all namespaces
> be represented as absolute URI.  That is a pointless waste of bandwidth
> and doesn't reflect the lessons learned from real-life usage of URI.

It might be 'real life' in the hypertext world but as we've already
found out it causes real interoperability problems when those
URIs escape their context....

> It is nice to be able to map any name to a universally-complete namespace with
> a standardized root, but only if the common case is for the important bits
> to be represented as a relative URI.  It doesn't even need to have a real
> base -- just an imaginary one.

It might be nice for you but it causes other applications no end of 
headache. Relative URIs should have always been defined as a semantic
of the HTML HREF tag and _never_ of URIs themselves. But those are
old arguments...

> Of course, nobody can do that with the URN syntax, which is why I won't
> be using it.

Which is the exact reason I use it and why I don't use HTTP schemes...

That's the nice thing about URIs. You use the scheme that works for
your application and I'll use the one that works for mine. Just as long as 
we don't try and start pushing application specific concepts into the 
whole URI space the world works just fine...


Michael Mealling	|      Vote Libertarian!       | urn:pin:1      |                              |

Received on Tuesday, 25 September 2001 22:36:47 UTC