[Top] [All Lists]

Re: [ontolog-forum] Current Semantic Web Layer Cake

To: "[ontolog-forum]" <ontolog-forum@xxxxxxxxxxxxxxxx>
From: Kingsley Idehen <kidehen@xxxxxxxxxxxxxx>
Date: Mon, 06 Aug 2007 11:14:25 -0400
Message-id: <46B73AD1.6070008@xxxxxxxxxxxxxx>
Frank Guerino wrote:
> Hi Danny,
> Some comments, below...
> On 8/5/07 4:22 AM, "Danny Ayers" <danny.ayers@xxxxxxxxx> wrote:
> >
> > We have something approaching standardisation on protocol (HTTP)
> HTTP is by no means the type of protocol that will allow "systems" to 
> communicate and interact with each other in such a way that will get 
> them to where they need to be for the vision of Semantic Web. HTTP is 
> a lower order protocol for simple put/get data to/from a system.
> > and document representation (HTML). There are still a large variety of
> > data representations, but many are becoming available as XML syntax.
> Again, this is also true with HTML, which is rudimentary, at best. For 
> systems to interact, there will have to be highly defined 
> dictionaries, grammars, languages, context analyzers, protocols of 
> interaction, and so much more. Think of how the brain works and how it 
> breaks down information and also how it uses it to categorize, store, 
> structure, correlate, index, find, recall, aggregate, transform, 
> format, render, etc. everything it works with. And, this does not even 
> get into the more complex aspects of how it takes all of this and 
> correlates it to motor skills, senses, communication with other 
> entities, etc.
> > The step to using a common baseline data model such as RDF may seem
> > improbable, but I personally believe that the gains it offers are
> > becoming visible enough to encourage widespread adoption.
> >
> > I think an important point to bear in mind is that 
> there_will_be_a_future!
> > There is everything to suggest that the future Web will involve more
> > connectivity between services and increased access to data that is
> > currently hidden in silos. There will be a "Web 3.0", how much energy
> > has to be spent getting there is another matter...
> >
> > The advantage of using RDF and associated techniques over arbitrary
> > data representations is that when a new service or repository is added
> > it's automatically part of the whole, without having to connecting to
> > a matrix of X^2 different representations.
> >
> > This doesn't mean every data/document producer has to publish RDF
> > directly, or that when they do it has to be hard work. Tools for
> > semi-automatically mapping from RDBMSs are available, and the GRDDL
> > technique [1] allows existing XML material to be viewed as RDF without
> > *any* additional work from individual publishers, just a one-off tweak
> > to the namespace document.
> Danny, I get you're point. However, take a look at what RDF is really 
> all about... It's about "Relationships". Most systems in the world 
> don't know how to capture, define, standardize, categorize, store, 
> communicate, etc. relationships. The world is still struggling with 
> the simplest of issues, which is dealing with flat content. In order 
> for RDF to really work, not only do systems have to speak the same 
> data dictionary language for flat content, they also now need to 
> understand how create, manage, and communicate "relationships". And, 
> anyone that has really tried to use RDF knows that it has many issues 
> that act as barriers for real use. If you want to use it in a 
> successful manner, you really can only use “pieces” of it that are 
> modified for real world use. We do much of this in our own system, 
> Danny. Trust me when I tell you that putting RDF to real use is “not” 
> an easy concept and a company like our own is an exception, not the 
> rule. Most enterprises in the world could care less about achieving 
> such a level of interaction because they're still trying to solve 
> business problems the traditional way. And, guess what... When it's 
> about making money, the traditional way is more than good enough and 
> probably will be for a very long time, making it highly improbable 
> that enterprises will switch to higher order standards.
Frank,    (01)

RDF Middleware can (and does) deal with the issue of producing and 
deploying RDF based Linked Data.    (02)

RDF (imho) standardizes the move towards making Concrete Conceptual Data 
Models the focal point of data interaction. Note, Apple achieved this 
eons ago with EOF (but it was landlocked in the NeXT realm). More 
recently, Microsoft has started to push the very same approach via 
ADO.vNext and initiatives such as Astoria Data Services (ultimately 
landlocked). I see RDF as the platform agnostic framework for 
interconnected Conceptual Data Models based on Web Architecture (no 
landlock whatsoever).    (03)

Today, you can convert non RDF into RDF Linked data (as stated earlier) 
via a number of middleware technologies:    (04)

1. GRDDL - for (X)HTML
2. Various RDFizers covering everything from SQL DBMS engines to 
spreadsheets and Web Services    (05)

The final piece of this puzzle is the deployment of RDF Linked data, 
this is a sticky issue because it ultimately redefines some of the 
features of the traditional Web Server. For instance, the Server has to 
be capable of the following:    (06)

1. Content Negotiation via HTTP
2. URL Rewriting
3. Pluggable Binding to RDFizers    (07)

I have written extensively about the "No RDF Tax" issue in a number of 
blog posts and demos over the last 12 months, I am also completing a 
white paper that provides a real life example of deploying SQL data as 
RDF Linked Data based on the popular Northwind SQL Database.    (08)

Putting RDF to use isn't easy (i.e. deploying as URI dereferencable 
Linked Data), but it is achievable with the right technology. RDF simply 
demands newer thinking re. server architecture.    (09)

--     (010)

Regards,    (011)

Kingsley Idehen       Weblog: http://www.openlinksw.com/blog/~kidehen
President & CEO 
OpenLink Software     Web: http://www.openlinksw.com    (012)

Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/  
Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/  
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/ 
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx    (013)

<Prev in Thread] Current Thread [Next in Thread>