Schemes like :

all attempt to re-invent the internet with something fixed, or something added.

But don't we already have the internet? Which already includes BillJoysSixWebs plus as many variations of medium, higher-level protocol and smart client?

The reason all these schemes don't look plausible is that we'll see all their properties evolve piecemeal, due to the WorldOfEnds property of the the current internet.

These alternatives will never take on an identity or an individuality of their own. They'll just be part of the internet. (And given internet means network or networks, it's own name seems sufficient.)

No grand design will replace the internet.


I keep reading people characterising the SemanticWeb as a "grand design" that is attempting to "re-invent the internet" - which, as you correctly point out, won't work.

Yet all the people I see /implementing/ the SemanticWeb don't have those aims. They see the SemanticWeb evolving piecemeal over time as you suggest.

All the people doing SemanticWeb work, writing the specs, etc. are implementing things BottomUp, and yet everybody criticises the SemanticWeb for being a TopDown design that won't work.

Odd that.


I see the point :-) In one sense, the SemanticWeb is no more centralized or TopDown than the original web was, because it required everyone to use the same language : HTML

So here are a couple of thoughts ...

First, this page isn't worrying about things being top-down, but about things not having a separate identity. But maybe that's wrong, too. Because, of course, everything that uses a special format or protocol is differentiable.

So there will clearly be a SemanticWeb ie. all those interlinked fragments of RDF. And I suppose there could be a KnowledgeWeb (as conceived by DannyHillis) consisting of some kind of information marked-up for intelligent tutoring systems.

Second, so why does the SW feel more centralized and oppressive?

I guess there are a couple of reasons :

1) It definitely is one-level of hierarchy / organization heavier than HTML or plain XML. Instead of creating an XML format for my data I need to have an RDF vocabulary defined by an RDF schema. As such vocabularies are normally intended as standards it seems like I can't start using it until I've got some committee of people together to agree on this standard.

(More discussion on SemanticWeb/SoHeavy)

2) It's just bloody difficult to understand and bloody difficult to do. It's so awkward, that it feels like something put together by a committees rather than by individuals producing something for their own purposes and convenience.

3) A lot of the hype about the SemanticWeb feels like a cult of TimBernersLee (The SemanticWeb, as brought to you by the W3C.) I guess this is just the media's way of understanding these thing so shouldn't really be held against those involved.

All right, so I'll quit bitching about the SemanticWeb ... but, you know if anyone can help with SemanticWeb/PleaseHelp then I'm eternally grateful.