[
Lists Home |
Date Index |
Thread Index
]
On 4/22/05, Michael Champion <michaelc.champion@gmail.com> wrote:
> On 4/22/05, Razvan MIHAIU <mihaiu@mihaiu.name> wrote:
>
> >
> > To be honest I do not understand what this article propagating. What
> > is the revolutionary thing ? If somebody has time, I ask him to drop me
> > a line or two about this.
>
> My question to anyone who attended: Why is the "simple single wire
> format for data" based on RSS/Atom rather than XML itself? Also, he
> keeps talking about a "web for data", but doesn't he mean what we
> usually call metadata?
It looks like these questions are answered in
http://www.onlamp.com/pub/a/onlamp/2005/04/22/bosworth.html
As I understand the argument: The problem with XML is that it isn't a
single simple sloppy syntax such as HTML is, it's a metaformat that
one can use to define many nice clean formats. Since everyone can
hard-code their knowledge of HTML (or RSS), they don't have to deal
with XML's meta-ness and can just get down to business. So, I guess
he's not talking about RSS as a metadata format to describe data, but
turning actual data inside out to fit it into the conceptual model of
RSS. Presumably the web would consist of documents in HTML and data
in this RSS-like format. That would let us query data using Google's
not-a-query-language rather than forcing us to use XQuery or SQL. We
wouldn't have to worry about schemas, or nasty joins, because
everything would have the same (basic) schema and be in flat
collections that didn't have to be joined.
This may play into the microformats idea, which I understand to mean
constraining existing generic formats with conventions for how to
address a specific domain. Contrast that wil the XML Way of defining
a specific syntax for each domain using the XML metalanguage.
I don't know what to say without sounding like Ted Nelson critiquing
the Web :-) I can turn my brain inside out and see how the Web might
evolve this way to accomodate data more pragmatically than it does
now. I can also see this turning the 'Net into an even worse sewer of
spam and malware than it is now, for roughly the reasons that Ken
North talks about earlier in this thread.
I dunno ... I can't say this vision appeals to me, but I can see the
momentum for RSS and microformats converging to produce this kind of
thing more easily than I can envision the Semantic Web, or WS-*
creating a worlwide services network, or XQuery as a way of integrated
XML views of diversely structured data on the web. The world would
probably be a better place if one or more of these came into being
rather than the Web of Data, but the world being the way it is, I
guess Bosworth's vision is a lot more likely to evolve, at least if
you buy Shirky's analysis in
http://www.shirky.com/writings/evolve.html
I don't have any trouble believing that schemas, XQuery, WS-*, maybe
even formal ontologies have a place in business and government
intranets, and the B2B world, and other places where orderliness has
a real payoff and rules can be enforced. But somebody PLEASE talk me
out of thinking that Gresham's Law makes Bosworth's vision all too
plausible for the Web itself: "bad" sloppy, simple but useful data
driving out "good" data that requires schemas or queries that are hard
to define and users that know the rules. Jonathan, you started this,
what do you think: ... will the XQuery paradigm or the Web of Data
paradigm prevail, or will they co-exist? [By the XQuery paradigm I
mean lots of diversely structured data integrated by queries on an
XML data model view that abstracts away the messy details ; by the Web
of Data paradigm I mean data messily stuffed into a generic schema,
with simple queries on the well-known structure].
|
