Thoughts about Radar Networks and Semantic Web

After reading this blog entry about Radar Networks and Semantic Web, here are some of my thoughts and follow-up questions:

  1. How will you cater to different type of users with different level of technical skills? (I guess you will provide different type of interfaces.)
  2. Are you seriously planning to stay in beta for 3 years? (Google is doing it. I guess you can find some ways to justify that.)
  3. How will you create enough incentives for end-users to spend the extra energy to “tag” or encode contents using your platform? (I am looking forward to see the beta.)
  4. Who will be in the position to define the ontology? Isn’t that usually done by a standard body? What if that standard body takes years to define such ontology? How do you regulate “the central ontology”? Will rest of the world agree with the ontology definition? What if a community of users wants to “fork” the definition? Will your platform offer interoperability across different silos of definitions?
  5. Although you are positioning Radar Networks as a non-competitor of Google by pointing out the differences of technical approach and focus, the actual end-user attraction factors might still related. Google is in the attention economy business. They focus on products and services that attracts attention from user’s eye balls, leverage that attention and convert it to ad revenue. One trick to attract eye balls is to present good enough search results to most users, effectively filter out a lot of noise on the Internet and present relevant information that users will most likely want to see. Semantic web is another creative approach to enable filtering massive amount of information and present relevant information to users. Although the technical approaches between Google and Radar Networks are different, the end effect that will attract users can be similar. Therefore, overlap of interest may still possible.
Advertisements

3 thoughts on “Thoughts about Radar Networks and Semantic Web

  1. Good questions. Although I’ve no affiliation with Nova’s company, I think I can answer a couple of the questions.

    1. Right – different UIs – your guess is as good as mine.

    2. Heh.

    3. End-user & extra energy – though I’ve no idea what Radara have got up their sleeve, there’s not necessarily any reason that users have to expend any more energy. One of the key ideas of the Semantic Web is better exploiting material that’s already out there. People already create lots of useful data – for example this blog is full of explicit data such as links, structured markup and there’s also the feed (there’s also all the stuff we have in our desktop apps and enterprise DBs). Where extra effort is needed, the answer is simply to make the benefits worth the effort 😉

    4. Another key idea of Semantic Web technologies is that there is no “central ontology”. The languages (RDF & OWL) allow anyone to make their own ontologies, the naming scheme of the web (URIs) being used to avoid clashes. But wherever possible existing ontologies should be reused, and/or connections made between them. For example, the FOAF ontology has stuff mostly for describing people, but it also has the fairly general term foaf:Agent (foaf:Person is an rdfs:subClassOf foaf:Agent). For fun I wanted to make a pet ontology, so I’ve got pet:Pet rdfs:subClassOf foaf:Agent. Now everything that applies to foaf:Agent also applies to pet:Pet. Any machine finding material using my ontology can follow the link to the FOAF ontology to get more information.

    5. Good point. I’d be surprised if Radar’s app didn’t feature some kind of search & filtering.

  2. There is a 4.0 example of “semantic information filtering” located at Boston Children’s Hospital’s “Center on Media and Child Health.” http://www.cmch.tv/research/. 4.0 NLP eliminates the need for “query structuring” by a user. 4.0 excels with lots of “context”- unrestricted length conversational style queries. Results are filtered and ordered more precisely when using “jargon”, well-articulated community-specific vocabulary. Try some “social sciences” questions like: What is the impact of the media on adolescent sexual attitudes and behaviors? Or, Can parents prevent children from experiencing unwanted effects of violent television?

  3. Danny, Thank you for attempting to answer my questions. Here are my responses/follow-up questions:
    3. Yes. That is exactly what I am trying to get the author to understand. Maybe folks at Radar Networks already understood this or maybe they just assumed that the platform will be so great that they might over looked it. I am not sure if they understand it from reading the blog. So, I asked the question.
    4. It’s good to know ontology also has the concept of inheritance. What if I also define my own version of pet ontology that is a sub class of FOAF and my definition of pet ontology has the following properties when compare to yours: 1. For the same resource, we may have similar definition and the variations on how we define such resource can make it really different from software/machine’s point of view? 2. For the same resource, we may have totally different meaning such that the definition actually contradict with each other. If the two properties that I described above are possible, then how can a third person effectively access both of our content with a single filtering, view, query or create a single consolidated ontology definition?

    Michael, Thanks for the example, which is using Semantx’s search solution. Semantx focus on special knowledge domains such as life science and health care, which makes the size of problem much smaller and simpler to solve.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s