W3C home > Mailing lists > Public > semantic-web@w3.org > April 2008

Re: discussion about Semantic Web realization

From: Danny Ayers <danny.ayers@gmail.com>
Date: Mon, 28 Apr 2008 18:03:27 +0200
Message-ID: <1f2ed5cd0804280903k62e7d06ctd87e77951bb29a9c@mail.gmail.com>
To: "Maciej Gawinecki" <mgawinecki@gmail.com>
Cc: semantic-web@w3.org
On 28/04/2008, Maciej Gawinecki <mgawinecki@gmail.com> wrote:
Nice set of points - not sure they're all covered by the FAQ...
http://www.w3.org/RDF/FAQ

The Semantic Web, in contrast:
> - requires a measure of centralization in order to make sense of
> schemas, i.e. the semantics cannot be built in to every client as the
> semantics of HTML and HTTP were built in to browsers


- it needs no more centralization than the current Web. The desired level of
semantics is up to the client/application, starting with links, e.g. the
Tabulator demonstrates that a generic data browser is feasible:
http://www.w3.org/2005/ajar/tab
http://en.wikipedia.org/wiki/Linked_Data

Triplestores are a prerequisite for a lot of data activities, and these
could be viewed as centralized off-web units (like many of the backend
databases behind sites/search engines). But a more useful perspective of
them IMHO is as caches of chunks of the Semantic Web - i.e. local views of
distributed data.

- requires much more cooperation from data sources (e.g. link targets)


Cooperation isn't essential, though reuse of existing vocabularies etc. can
help maximise the benefits of the network effect.

- is based on a complex markup (RDF) that's difficult for
> non-programmers to work with


RDF/XML is notoriously hard work, but there are alternatives like Turtle and
RDFa :
http://en.wikipedia.org/wiki/Turtle_%28syntax%29
http://www.w3.org/TR/xhtml-rdfa-primer/

It should be remembered that RDF is essentially a data model, not a markup
format. The conceptual model can be understood by a six year-old:
http://www.ldodds.com/blog/archives/000329.html

Also (like HTML), the aim isn't for end-users to work with the code
directly, but via tools which are appropriate for their needs/skills. The
data from most existing applications can be expressed as RDF without any
change in behaviour by the user (assuming there's a developer of two around
to do the mapping...).

- has to compete with its predecessor and many other technologies


The Semantic Web is an extension of the current Web, in particular relations
in RDF can be seen as an evolution of the link:
dsonline.computer.org/portal/pages/dsonline/2007/06/w3web.html

- has very little commercial interest, unclear revenue model, etc.


There is interest from many different sectors & companies - e.g. Microsoft,
IBM, Google and Yahoo! have all got something going on in the field. More
telling perhaps is the existence of a business-oriented conference:
http://www.semantic-conference.com/

The revenue models will vary from case to case. Patterns may well emerge
that aren't yet clear - just like it's been going with the Web from the
start.

- requires a critical mass of participating sites to be interesting and
> useful


The technologies are already demonstrably useful (e.g. I'm now using Twine
daily), though the really big gains will only start emerging once there is
wider participation - the network effect again.

Most importantly, the development of the World Wide Web was driven by
> the need to solve real, primarily technical problems. The Semantic Web
> was also invented in order to address a real problem (sorting through
> information) but because of the reasons above and the existence of
> alternative solutions (e.g. Google) the same driving need has failed to
> materialize.


The problem is the same - the document Web only gets us part of the way,
support for data in general will improve and expand its capabilities. Google
is a partial workaround for the fact that a lot of the useful information on
the Web isn't available for direct machine processing, it doesn't really
help the underlying problem. The document Web isn't going away, and there's
no requirement that every Web publisher has to publish data to benefit from
the data web - an inkling of the potential there is visible in community
tagging sites. But any increased exposure of data that *already exists* in
the world's databases to the web (in an interoperable fashion) is going to
make a big difference.

The Semantic Web is a classic "second system"
> (http://en.wikipedia.org/wiki/Second-system_effect).


The Semantic Web is just a phase the Web is going through :-)

Cheers,
Danny.

-- 
http://dannyayers.com
~
http://blogs.talis.com/nodalities/this_weeks_semantic_web/
Received on Monday, 28 April 2008 16:04:07 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 1 March 2016 07:42:04 UTC