[
Lists Home |
Date Index |
Thread Index
]
Paul T wrote:
>
> I guess that what was implemented was to make a validator
> to fetch the schema from RDDL Place, when switching from
> one namespace to another or something. The implications of this
> are so *huge*, that I would not even start explaining why it is a
> bad thing to do. So it is no longer 'just documentation with a
> few utility links in it'. It is a hidden extension to schemas, for
> example.
>
My idea for RDDL is that when a namespace qualified XML element is
encountered, by either a person or a computer, that one can, assuming the
namespace URI referenced a RDDL document, discover some useful information
regarding how the element is intended to be processed.
Assuming the encounter is with a human, I imagine that it would be desirable
to read a natural language description of the element, hence XHTML.
Assuming the element is encountered in the process of schema validation, I
imagine that it would be desirable to locate a schema suitable to validate
the element.
Assuming that rather than an XML element, a QName is encountered as the name
of an XPath function: it is desirable to find an appropriate piece of code
which can execute the function.
So yes, RDDL might be a way to locate Netscape plugins, or IE objects. The
point is that the same element might point to all of a netscape plugin, IE
COM object, Hotjava applet etc. The creator of the namespace can place any
number of pointers to 'plugins' for various browser architectures. (this is
one example).
RDDL can be used as a way to create distributed java applications (via the
RDDLClassLoader).
I agree that the implications are huge. You cannot start to explain why this
is a bad thing to do. I consider this a good thing, indeed a basic part of
some form of the 'Semantic Web', but I would never try to force you to do
things this way. If you consider this a 'bad thing' and I consider this a
'good thing' then so be it.
Jonathan
|
