ReadWriteWeb via Alex
Iskold have delivered another iteration of their "Guide to
Semantic Technologies".
If you look at the title of this post (and their article) they seem to be accurately
providing a guide to Semantic Technologies, so no qualms there. If
on the other hand, this is supposed to he a guide to the "Semantic Web" as prescribed by TimBL then they are completely missing the
essence of the whole subject, and demonstrably so I may add, since
the entities: "ReadWriteWeb" and "Alex
Iskold" are only describable today via the attributes of the
documents they publish i.e their respective blogs and hosted
blog posts.
Preoccupation with Literal objects as describe above, implies we
can only take what "ReadWriteWeb" and "Alex Iskold" say
"Literally" (grep, regex, and XPath/Xquery are the only tools for searching
deeper in this Literal realm), we have no sense of what makes them
tick or where they come from, no history (bar "About Page" blurb),
no data connections
beyond anchored text (more pointers to opaque data sources) in post
and blogrolls. The only connection between this post and them is
the my deliberate use of the same literal text in the Title of this
post.
TimBL's vision as
espoused via the "Semantic Web" vision
is about the production, consumption, and sharing of Data Objects
via HTTP based Identifiers called URIs/IRIs (Hyperdata Links / Linked Data). It's how we use the Web as a Distributed Database where (as Jim Hendler once stated with immense
clarity): I can point to records (entity instances) in your database (aka
Data Space) from mine. Which is to say that
if we can all point to data entities/objects (not just data
entities of type "Document") using these Location, Value, and
Structure independent Object Identifiers (courtesy of HTTP) we end
up with a much more powerful Web, and one that is closer to the
"Federated and Open" nature of the Web.
As I stated in a prior post, if you or your platform of choice
aren't producing de-referencable URIs for your data objects, you
may be Semantic (this data model predates the Web), but there is no
"World Wide Web" in what you are doing.
What are the Benefits of the Semantic Web?
Consumer -
"Discovery of relevant things" and be being "Discovered by relevant
things" (people, places, events, and other things)
Enterprise - ditto
plus the addition of enterprise domain specific things such as
market opportunities, product portfolios, human resources,
partners, customers, competitors, co-opetitors, acquisition
targets, new regulation etc..)
Simple demo:
I am a Kingsley Idehen, a Person who authors
this weblog. I also share bookmarks gathered
over the years across an array of subjects via my bookmark data space. I also subscribe
to a number of RSS/Atom/RDF feeds, which I share via my feeds
subscription data space. Of course,
all of these data sources have Tags which are collectively exposed
via my weblog tag-cloud, feeds subscriptions
tag-cloud, and bookmarks tag-cloud data spaces.
As I don't like repeating myself, and I hate wasting my time or
the time of others, I simply share my
Data Space (a collection of all of my purpose specific data
spaces) via the Web so that others (friends, family, employees,
partners, customers, project collaborators, competitors,
co-opetitors etc.) can can intentionally or serendipitously
discover relevant data en route to creating new information (perspectives) that is
hopefully exposed others via the Web.
Bottom-line, the Semantic Web is about adding the missing "Open
Data Access & Connectivity" feature to the current Document Web
(we have to beyond regex,
grep, xpath, xquery, full text search, and other literal
scrapping approaches). The Linked Data Web of de-referencable data object URIs is
the critical foundation layer that makes this feasible.
Remember, It's not about "Applications" it's about Data and
actually freeing Data from the "tyranny of Applications".
Unfortunately, application inadvertently always create silos (esp.
on the Web) since entity data modeling, open
data access, and other database technology realm matters, remain of
secondary interest to many application developers.
Final comment, RDF facilitates Linked Data on the Web, but all
RDF isn't endowed with de-referencable URIs (a major source of
confusion and misunderstanding). Thus, you can have RDF Data Source
Providers that simply project RDF data silos via Web Services APIs
if RDF output emanating from a Web Service doesn't provide
out-bound pathways to other data via de-referencable URIs. Of
course the same also applies to Widgets that present you with all
the things they've discovered without exposing de-referencable URIs
for each item.
BTW - my final comments above aren't in anyway incongruent with
devising successful business models for the Web. As you may or may
not know, OpenLink is not only a major platform provider for the
Semantic Web (expressed in our UDA, Virtuoso, OpenLink Data Spaces, and OAT products), we
are also actively seeding Semantic Web (tribe: Linked Data of
course) startups. For instance, Zitgist, which now has Mike Bergman as it's CEO alongside
Frederick
Giasson as CTO. Of course, I cannot do Zitgist justice via a footnote in a
blog post, so I will
expand further in a separate post.
Additional information about
this blog post:
- I didn't spent hours looking for URIs used in my
hyperlinks
- The post is best viewed via an RDF Linked Data aware user
agents (OpenLink RDF Browser, Zitgist Data
Viewer, DISCO Hyperdata Browser, Tabulator).