Re: Hyperlinks depend on GET (was: Re: REST and the Web)

On Saturday 30 March 2002 03:16 am, Paul Prescod wrote:
> You don't need GET in closed-system hyperlinking. But I was talking
> about the Web and I didn't expect that the statement would be
> applied to another context.

Well "without GET you don't have hyperlinking." is pretty much 
hyperbole.... this is only true in HTTP and only true in the WWW as we 
see it now.  

> I didn't say that GET gives you everything you'll ever want from
> hyperlinking. I said that you need to be able to GET data to
> traverse hyperlinks (to data you don't already have).

Personally, I don't think GET captures the notion of link traversal 
well at all (one indication being the use of fragment ids). If the WWW 
were cast as a large-scale DSM, for example, the idea of "fetching" 
(GET) becomes a less accurate modelling of the semantics of the 
underlying system model.

In addition, hyperlinking isn't dependent on/defined by traversal. 
Hyperlinking is dependent only on link specification/discovery. Links 
and link traversal are logically distinct.


Received on Saturday, 30 March 2002 13:19:40 UTC