Lists Home |
Date Index |
> From: Paul T [mailto:firstname.lastname@example.org]
> Is RDDL for 'computers' or for 'humans' ? What is the
> real problem RDDL solves and who is using RDDL?
It's for both computers and humans. It's human readable documentation, with
embedded machine readable links. When an application encounters an element
in a particular namespace, if RDDL is at the end of the namespace, then the
application can easily locate a schema that can be used for validation,
reference documentation, stylesheets for displaying or converting to another
format, or just about about any other sort of associated resource that may
The prevailing model of relying upon an instance containing a
"schemaLocation" attribute to locate a schema, and an "xml-stylesheet" to
locate an appropriate stylesheet, and other associated PIs, attributes,
and/or links for each and every relevant resource has obvious problems.
The use of XLink in RDDL allows RDDL to be agnostic with regard to
particular resource types. RDDL does not need to know the difference between
a schema and a stylesheet (or any other type of resource that someone
invents). They are just resources with a distinct "nature" identified by a
URI (typically the namespace of the document element). Only the application
that wants to use a particular type of resource needs to understand that
type of resource, and it identifies the desired type of resource via the
"role" (and optionally "arcrole") attributes on the link. It's a bit
abstract, but very flexible and extensible. If someone invents a new type of
resource that may be useful, RDDL already supports it without changes.
> To me, RDDL is just a nice looking paper, responding to
> the fact that there 'could be something at the end of
> namespace URI'.
> Problem 'what will be at the and of namespace URI'
> is still neither solved, nor addressed and it is also
> orthogonal to 'how can we make XML self-describing'.
I'm steering clear of that whole "self-describing" thread. I agree with
Michael Kay's remark on this. The whole self-describing thing just leads to
unresolvable paradoxes. (I think it was Michael Kay who posted that. I can't
find the post, now.) RDDL can, however, permit an application to locate
schemas and human readable documentation for an XML grammar based on its
namespace. I think that is about the best we can get.
> The explanations are in my previous letters on this subject.
> I believe that it would be better for me to disappear from
> this thread now.
Well, I won't try to keep you in this thread any further, but I could not
resist making a quick comment on this.