[microformats-discuss] FYI: two posting about the Semantic Web, the "SynWeb", scraping and microformats

[Danny Ayers]

On 10/24/05, Tantek Çelik <tantek at cs.stanford.edu> wrote:
> On 10/24/05 9:43 AM, "Danny Ayers" <danny.ayers at gmail.com> wrote:
>
> > [1] http://dannyayers.com/2005/09/cake.gif
>
> There are just soooooo many things wrong with that diagram that it's hard to
> know where to start criticizing.  In fact, not sure it is worth criticizing,
> so I'll just point out the only pieces that I think make at least some
> amount of practical sense:
>
> * URI (and even then I'm starting to dismiss URNs are mostly academic, thus
> leaving only URLs),
> * Unicode (I admit, because of my last name I'm biased here)
> * XML (a reasonable foundation, even if XML+tidy works better in practice)

I can't deny that diagram has huge flaws, it's not only old but I
think it was mostly handwaving in the first place, you shouldn't take
it too literally (I could be wrong - check the design issues [1]).  It
lacks reference to the protocol(s) for a start. XML Schema is a pain.
But a lot of it works. The key notion of layering works, e.g. if you
only use URIs+XML it still makes sense from above. RDF+S and ontology
(OWL) layers have been well covered, the logic layer (to my eyes at
least) seems to have merged itself down a level or two. Proof has
implementations, but its utility in the distributed environment has
yet to be demonstrated. The digital signature bit is still in
neverland, as is the trust on top. There are two big sidepieces that
weren't really envisioned when the diagram was drawn up - GRDDL and
SPARQL, either of which could prove paradigm-shifters.  My own
revision of the diagram is at [2].

> The rest are best left to academics IMHO.

They're the last people I'd want to leave the future of the Web.

Personally I think the generalising of the Web from a poorly indexed
document repository into a more general database, communication and
computational unit is what should (for everyone's benefit) and what
will (what else?) happen. The Web as platform and all that.

At this point in time the existing protocols are pretty capable, the
basis for comms is in place but services tend to be a bit clunky,
suggesting they need a different approach to wiring, a more
declarative (data-driven) approach seems the most promising.

The RDF model seems to have relatively low friction when it comes to
bringing in the data, easily mappable to RDBMSs, objects and XML
formats (e.g. microformats can be transparently covered). The logical
model carries virtually all the same attributes of the existing Web
(shifted up a notch conceptually), which suggests it can work at Web
scale.

What happens on the desktop/mobile is another question - maybe
Spotlight and WinFS give hints.

What kind of path ahead do you see for the Web in the next few years?

Cheers,
Danny.

[1] http://www.w3.org/DesignIssues/Overview.html
[2] http://semtext.org/2004-02/slides/img4.html

--

http://dannyayers.com