Re: "State of the Semantic Web" - personal opinions?

Hi Danny,

Will share about my company SemanticSoft, Inc tools. We finally started 
treating SemanticWeb primarily as a “world wide data bus”, as Tim put 
it. Our second vision-shift was that SemanticWeb technologies should 
move “behind the scenes”, so that the users see a “classic UI" which 
does not require knowledge of standards. A good return on investments 
can be obtained from “regular” tools equipped with SemanticWeb 
technologies "in the background" to make the tools interoperable, 
flexible and mobile. I started treating SemanticWeb as value-added 
"regular" software.

We are currently changing our tools described here 
http://www.semanticsoft.net/semanticwebtools.html towards by hiding 
SemanticWeb behind a customary UI.  You can see the *current* versions 
by clicking "Try" against a tool on our site - one standalone tool can 
be downloaded, with other two web based tools you get from the site into 
our servers where you can do work specific to them.

*SemanticServer* is a “gateway” to SemanticWeb. If you click on Try, you 
get to a “desktop” with a Start button in corner like in Windows. But 
you are on the web and you can manage resources and their metadata in 
any vocabulary. The Help file (65 pages) of "Start button" explains what 
you can do in this space. You can add new "hosts” to manage resources in 
repositories on other instances of SemanticServer, or any other 
repositories compliant with JCR, Java Content Repository, specification 
JSR-170. We chose JCR because this appears to be the only standard for 
content repositories for CMSs. You can search resources by our built-in  
*visual* SPARQL engine. You can also search *inside* some resources - 
the server stores resources which are not blobs in such a manner that it 
can "see" inside them. This tool could be used in corporate space, BUT I 
think, like many other tools on SemanticWeb, which require knowledge of 
the standards it will be *not* be used. With next version we hide 
SemanticWeb technologies to make out of it a "regular" but intelligent tool.

Our new tool *ResourceDescriptor* will be used both as built in 
SemanticServer, or as a separate tool for those who want to use it 
outside JCR "space".  ResourceDescriptor can be treated as an 
"intelligent client" which has wizards so that a user without any 
knowledge of Semantic Web can write any description or make a  SPARQL 
query. When this tool is released and built into SemanticServer, I 
think, we get be a tool for "corporate space", or "content business" 
like eBay,  Amazon.com, etc.

The version "in works" of  *SemanticStudio*, current version of which 
you can see the web site, will also have a "Human Interface" like in 
ResourceDescriptor. In the autumn it will have an "object oriented" 
graphical UI like in Rationale Rose. All other features remain as 
described on our site. When next version of S-Studio is ready, it can be 
used also by people who do not know the standards. When the graphic mode 
will be ready the people using UML will be able to work in an 
environment almost familiar, but will save the project as an ontology.

"Chameleon" project illustrates our idea of a "semantic" web site. It 
will be a CMS, but will import/export web content formatted in 
SemanticWeb standards, and this will allow to exchange web content with 
other instances of Chameleon or other tools which can work with 
OWL-formatted web content. You will be able to save a compartment (many 
linked pages and users, usergroups, and their access rights) of a site 
and send it to another location where you can import it into a location 
so that the pages are automatically built, with or without 
users/usergroups/accessRights, etc.  Currently, Chameleon is a regular 
CMS and you can get to it and play directly from our site. Unfortunately 
we don't have money to complete, and we stopped implementation of this 
idea until better times. I just wanted to share our idea a "semantic site".

>
> * Money! What's the current status of funding for semweb research in 
> academia? Inside big corps? Gov. orgs? Funding from VCs etc?
In Moldova, where we are located, there is little of this :-). We 
finance our SemanticWeb development from small profit from "classic" 
software they order to us. Big foreign corps are outsourcing to Moldova 
only development of "classic" software. Now we are waiting for a big 
corp which would like to build SemanticWeb tools cheaper...
>
> * What's the range of application of RDF like nowadays? (Obscure 
> examples would be nice)
1. *Data reconciliation* is the strength of RDF and no other standard 
can compete on this. Data coming from different sources in different 
formats (and even different media types) can be reconciled by conversion 
to RDF. The biggest impact would be on B2B. I am coming to SemanticWeb 
from EDI - the "backbone of e-commerce". The situation in B2B is like 
this - businesses doing B2B are grouped in a couple dozen "closed 
worlds" each governed by one standard. To exchange data between such 
worlds is almost impossible, because each standard is complex and 
standard-to-standard conversions are hard to implement. If nothing is 
done, the EDI will become next "millennium bug". There is a rather 
simple idea how to reconcile the standards based on RDF, but I will not 
share it here since it would take some technical details.

2. In sound recognition, the methods based on physical characteristics 
are not sufficient even for music/speech discrimination in soundtrack. 
Neural networks raised the quality close to "speech recognition", but 
without continual "guesswork" and "anticipation" the quality is still 
low. There are strong relationships between RDF graphs and neural 
networks (roughly, a logical "pattern" in speech is a RDF graph).  
Therefore RDF can serve as the *logical layer* in speech recognition. 
Intellectual Property domain is the main investor in this area and IP in 
entertainment industry probably is among largest. They would value a 
program to get some control over millions of Radio/TV channels, in order 
to protect IP of pieces of music, films, etc. Also, Radio/TV pays big 
money for "simple" programs of music/speech discrimination needed for 
statistics on radio/TV. I am sure that protection of IP and 
statistics/planning in Radio/TV will be next great application of RDF. 
We gave good thought to this, because some of our developers formerly 
worked on such a program for TV and now they see how they could have 
done this work better, if they knew  RDF at that time. Now we are 
looking for such an order to try this technology.
> * Has the role of the W3C changed in this context over the past few years?
The W3C's help on SemanticWeb to companies is really great. I wonder how 
the SemanticWeb activities leads manage to read and reply in detail even 
to a message addressed personally. I believe,  W3C changed from a 
standards organization to also a *dev lead* on large scale projects. But 
there is also a dark side to this - now investors are asking why you 
(W3C) did not yet complete the project :-)   
>
> * Slightly tangential - where do you see social networking going? 
> (Possibilities off the top of my head - unification of services; 
> general loss of interest through another fad coming along; descent 
> into the Web infrastructure) 
> - supplemental: assume the fad prognosis - what'll be next?
I believe, *descent into the Web infrastructure* is the correct path to take
>
> * Not unrelated, there's a fair bit of similarity between OpenID 
> Attribute Exchange and RDF, as well as what appears to be a parallel 
> stack to the (Semantic) Web with XRDS/XRI/XDI etc. Is independent 
> invention of this nature a good thing or not?
> (My mouth remains firmly shut :-)
>
I think, everything fits or can be adjusted to fit into URI as a 
*scheme*.  I believe, a good thing are only inventions which fit in the 
SemanticWeb project stack.
>
> * Jim Hendler's question: where are the agents?

Suppose we complete our Chameleon "semantic site" and it saves itself 
into a file and sends it to other compliant location, and in the file 
there are "new" users with access rights, then this content get into 
other sites with "controlled access"... there are lots of scenarios 
which make such agents worse than global warming.

I believe in direction of intelligent agents  lies the "semantic virtual 
machine" which I outlined in my article "A Wholebrain Approach to the 
Web" at "Web Intelligence - Intelligent Agent Technologies", Silicon 
Valley, Nov 2-5.
When SemanticSoft become richer :-), we are planning to implement this.

Ioachim Drugus, Ph.D
Main Architect,
SemanticSoft, Inc
http://www.semanticsoft.net

Received on Wednesday, 2 July 2008 22:37:06 UTC