Lists Home |
Date Index |
Nicolas Lehuen wrote:
> Well, I hope I won't be as flamed as Paul, but I have a few remarks too
> make me reluctant to use RDDL.
And I hope that I would never flame anyone for asking a specific question,
nor for making a specific suggestion.
> 1) Scalability. With only a few arcroles (or purposes ?), a single RDDL
> document is readable. What if the number of associated resource grows ?
> is the benefit for a human reader to have to download a 500K document to
> have a description of ...
That might be a problem if you were to have a huge namespace. Then again
your schemas would also be huge, so efficiency would be a concern regardless
> 2) ... a description of what ? A namespace ? If RDDL describes a
> then let's be careful when providing arcroles for DTD and schemas. The
> where one namespace = 1 DTD and 1 schema document for each schema language
> that you want to support is, like Paul wrote, a dangerous degradation of
> namespace purpose.
So are you saying that a nature and a purpose is not enough? Do we need 3
URIs to describe a resource. What axis is the third URI? Do you have an
example. Would such an example be common?
I don't have any authority on the subject (and I'm sure
> you people won't miss the opportunity to prove it to me), but I firmly
> believe that the interest of namespace resides in allowing tags from
> different namespaces to be mixed in the same document without name
> collisions. If I'm supposed to use RDDL to find the proper schemas or
> stylesheets or whatever, where can I find a RDDL document for this XHTML
> document that contains tags from the MathML namespace ?
Excellent question. Henry Thompson has included such RDDL support in the XSV
At the XHTML
> namespace URL, or at the MathML namespace URL ? It's funny to notice that
> the same problem exist for RDDL document themselves, since they mix the
> (X)HTML, RDDL and XLink namespaces... The problem is that with
> namespace-centric view of RDDL just prevents document types that can mix
> tags from different namespace from having an associated RDDL document (if
> doesn't, show me the URL where I can find it), resulting in a big hole in
> the practical usability of RDDL.
No, I believe Henry's implementation demonstrates that RDDL can be used in
such multinamespace situations. If RDDL were not useful in such situations,
I agree that this would be a serious limitation.
> 3) It is a bit related to the scalability issue, but how do you handle
> internationalisation ? RDDL contains human-readable text, that's fine, but
> not everybody can or want to read English. So will you have all possible
> translation of the human-readable text in the SAME RDDL document ?
RDDL does provide for internationalization via the xml:lang attribute. I
would not claim that this solves all internationalization problems, then
again what happens when a French browser hits a German page (there must be a
punch line in there somewhere). RDDL is certainly no worse than the basic
HTML/HTTP Web infrastructure in this case.