Dedicated to Bob Abooey,
the true Godfather of the Funk
- One of the users of RSS
Bandit took the capability to create custom
themes via XSLT to the next level and his custom
theme uses Javascript to create a voice enabled
hottie who reads blog posts him. Here's a
screenshot of the
Living Actor enabled RSS Bandit. You should
visit his
site (Warning: Lots of active content)
-
Matrix Reloaded: The Abridged Script
- Fumiaki Yoshimitsu
still doesn't see anything wrong with
derivation by restriction.
So, my concern right
now is whether it is really so problematic to
validate DBR-ed instances and schemas using
different validaters. But Dare, if it is so,
then what sucks is not DBR but
validaters...Are they working so hard, or the
spec is so sucked they are not willing
to?
Part of me really wants to take offence at
this post but it's not Fumiaki's fault. I
remember being in college and thinking to
myself "How hard is it for these bonehead
companies to implement a simple spec?" Now that
I've actually worked closely with some of these
specs my opinion has significantly changed. I'm
not going to bother trying to point out how
complex the XSD spec is because you can just read
it yourself. In fact, I can ask a few
simple questions to show how straightforward
the spec is
- What are the differences between block
and final on elements vs. complex types?
- Can a simple type be derived from a
complex type?
- Can xsi:schemaLocation appear anywhere in
an instance document? If so what are its
effects on previously validated
elements?
- Does XSD disallow non-deterministic
schemas? If so what rules in the spec
actually disallow them?
- what do processContents="lax" and
namespace="##other" really mean on a
wildcard?
These are five straightforward questions not
tricky ones by XSD standards (because those
involve actually writing out schemas).
Attempting to correctly answer such questions
using the various specs as a guide is
instructive and should help one understand how
difficult the folks writing XSD validators have
it.
My thinking is that DBR is for local
extension, while open content model is for
outsiders. If an XSD schema was created by
myself, then I can use DBR for my next
schema. If an XSD schema was created by
others, then I can't extend the schema unless
they have defined xs:any in their schema.
The above statement doesn't make sense to me.
Derivation by restriction means the subtype has
a content model that is a subset of the
content model of the base type so I fail to see
how one can use this for local
extension. Secondly one can derive from
types authored by other parties and in fact can
use substitution groups or xsi:type in
instances to ensure that the derived types are
used in place of the base types where they are
expected in an instance
- I bought a TiVo this weekend.
I love it already although I have to teach it
that just because I like X-Men:Evolution and
Justice League doesn't mean it should
preemptively record every brain dead early
morning cartoon show.
- A little while ago Ray Ozzie
contacted me about some issues he was having with
RSS Bandit which mostly boiled down to the fact
that it does not process feeds that contain
invalid XML even though this seems to be standard
practice amongst RSS clients. Both Torsten and I
firmly think this is a bad practice.
Dare's Opinion
Torsten's Opinion
-
Get yourself a
News Aggregator and subscribe to my
RSS feed Disclaimer:
The above comments do not
represent the thoughts, intentions, plans or
strategies of my employer. They are solely my
opinion.