A conversation with Yihong Ding

Yihong Ding is a researcher in the realm of Semantic Web. He is also a contributing author of SemanticFocus.com, which is one of the most active blogs around the topic of Semantic Web. Because of the many commons shared between me and Yihong, I decided to write to him about a specific direction of my research. Suprisingly, I had his reply not even after 20 mins. In his email, he patiently and expertly explained his thoughts about current problems of semantic web again followed by a couple of brilliant suggestions in order to answer my question.

Here is the conversation:

Tao:

Semantic Web Service (SWS) has been a round for a while. and I believe you must had a lot of understanding and research in this area. As you might know, researches in SWS are mainly focusing on semantic service description such as WSDL-S. However, the importance having communication payloads such as SOAP equipped with semantics has been somehow ignored. Occasionally, I found the paper " Combining RDF and OWL with SOAP for Semantic Web Services" has pretty much the same idea. But that was published on 2004, and I failed to contact the author, though he is a Chinese scholar as well. Personally, I think RPC style WS has its deficiencies to enable more flexible business funtionalities due to the issues such as versioning. Message-driven web service possibly with semantics would be a good choice to combine computational power in order to provide some more flexible or even some intelligent business functionalities. But my concern is why I can't find any more papers to support this assumption.

Since you have much more in-depth understanding of Semantic Web, would you be able to give me your view on this topic? What would be the potential problem of having semantic soap? I'm really looking forward to have your reply. And again, it's a great pleasure to know you, hopefully we can have more discussions later on.

Yihong:

Although semantic web has been discussed and studied for several years, some fundamental issues about realizing semantic web is still open questions. One of the reason is due to the progress of web evolution. The jump from traditional Web 1.0 directly to the ideal Semantic Web is too far away. This was what Tim Berners-Lee had not thought when he wrote his book at 1999 and his Sci-American article at 2001. With the gradual mature of Web 2.0, the realization of semantic web start to being possible. But still, there are fundamental issues we must understand before the dream could come true. This is thus the motivation of the post you read and several other posts I have written both at SemanticFocus and my own blog.

Back to your question. I am sorry that I am not an expert on the web service realm. So I am not the best one to answer your question. But I can try to tell you what I think of the problem is.

The problem of SWS still lays on the problem of sharing semantics, which is fundamental to semantic web. In fact, particular language itself (such as RDF or OWL or WSDL-S or any other ones) is not the core of this problem. The problem is how to share. Comparing to regular texts, services has left less spaces for people to specify their semantics. Therefore, it leaves even less spaces for people to share semantics.

What I can suggest to you is that probably you can watch more about web service architecture literatures before you dig deeper into the realm of SWS. The current real problem is a better architecture. Specific SWS languages could be efficient only if they are well coped to the architecture. Hopefully these suggestions could do some help.