RE: Update on namespaces

At 16:08 1997 06 10 -0700, Andrew Layman wrote:
>I indeed think of types as equivalent to notations, as your mail points
>out. Looking at notations, we see that they identify the format of an
>external binary entity, and associate that format with a public
>identifier.  That is, notations do specify data types, and within the
>limits of XML public identifiers, notations are universally unique. 

>
>However, notations are not the same as namespaces. Namespaces are not
>data types. Namespaces are a general mechanism that allows any element
>name (or by extension, any name) to be associated with a particular kind
>of system literal, a URI, and thereby namespaces allow any name to be
>made universally unique. Notations contain a mechanism by which
>notations can be universally disambiguated; but it only works for
>notations, it isn't general.

I might start an avalanche down on my own head, but I'm not sure I
agree that notations are so narrowly defined.  In fact, I'm not sure
notations are so well-defined at all.  I do not see why a notation name
couldn't refer to a namespace.  

Certainly a notation declaration associates a notation name to any system 
literal including a URI--that seems unarguable from merely a syntatic 
viewpoint--and I don't see why one couldn't use this mechanism to make a 
namespace id to a URI that defines that namespace.

paul

Received on Tuesday, 10 June 1997 22:40:00 UTC