RE: [xml-dev] Meta-somethingorother (was the semantic web mega-permathread thing)

["Joshua Allen" <joshuaa@microsoft.com>]

> controlled environments, multiply it by n, and jam it all together
> somehow, then <poof/> some emergent magic happens and we have
something
> wonderful and new.

Like Doctorow with metadata, you are projecting your own ideas onto what
I said.  I have said nothing about such a utopia.  I simply pointed out
that having your metadata stored in a triples format can be very
*useful*.  Surely you cannot be arguing in favor of the converse.

Now, there could be some argument over *how* useful that will be.  I'm
in no way claiming a utopia, but I believe it will be quite significant.
You seem to be talking "emergence" and "AI", when I am simply talking
economies of scale and interop.

> I'm sorry Joshua, but after decades of failure in this particular
strand
> of AI, the burden of proof is on the advocates of this dream not the

Again, I don't know where you read "AI" in anything I said.  I believe
you are again projecting your own internal utopia onto what I said.  I
believe that "semantic web" could be incredibly useful without ever
performing a single inference.

> skeptics. Come back when you've got some evidence not just hand waving
> and promissory notes.

I remain open minded about how realistic (or not) the vision is.  I am
just trying to explain *what* the vision is, since so many people seem
to be projecting their own erroneous assumptions onto "semantic web".
Two of the most egregious assumptions that people falsely project are:

* it's primarily about the "META" tag
* it's about flawless information and AI, agents, etc.

I think that both of these assumptions miss the 90% value proposition
for "open metadata".