Bernie Sloan wrote:
> Jim Weinheimer said:
>
> "I think it may still become a basic focal point for the Semantic
> Web."
>
> Feel free to smack me if this is a dumb/naive question, but how would this be a
> basic focal point for the Semantic Web?
>
> I'm not doubting the statement...I'm just trying to understand it.
>
> Bernie Sloan
This always scares me. I remember the saying of Mark Twain, "It is better to keep your mouth shut and let everybody suspect you are a fool, than open it and remove all doubt." So, I'm opening my mouth. I just hope my understanding isn't too far off the mark.
I have my own understanding of the Semantic Web. What Eric wrote, I agree with, but I would just like to add a few points. The way I see it, there are essentially two areas:
Let's say you are looking for a new job on the web (just like I’m sure many of us are doing right now), and have you seen those new sites lately? You have your beautiful resume all done, but they want you to reproduce it in their system so that they can run through it quickly in their database to save their own time and energy. So, you have to sit there, copying and pasting, or retyping the same things in, over and over and over, for each job. I understand their needs, but isn’t there an easier way?
The Semantic Web offers a method that, if you code everything correctly in your resume, it would automatically feed into their system. Being librarians, we understand that this is similar to bringing all the world’s library catalog MARC fields together, so that 100/700 in MARC21 equals the equivalent fields for personal authors in Russian MARC, the equivalent in other MARCs and so on. If we could get to that level, it would make it much easier to share data.
As librarians, it is natural to ask, "It's nice to link fields, but what about the forms of the names, e.g. Samuel Clemens or Mark Twain?" This is the other part I mentioned, or the "conceptual part" that Eric wrote of. In essence, that any concept will be expressed as a URI, not by a textual label. Therefore, the idea is that if URIs exist, everybody can link to the same concept, e.g, I have a web page, where I write:
"I love my cat."
When Google eats it, people can find my page when they search "cat."
But with the correct coding I can do it "semantically":
I love my
<item rdf:about="http://dbpedia.org/resource/Cat">Cat</item>.
This would allow people to link all "cat" values together, and these could be used for searching. The word "Cat" may appear as Gatto, Kot, or whatever the user wants.
Click on that that link to dbpedia and see how it works. In a semantic web search engine, you could search for "cat" and it would actually search for the URI, in this case, http://dbpedia.org/resource/Cat, and bring things together in this way.
Another possibility:
<item rdf:about="http://dbpedia.org/resource/JimWeinheimer">I</item>
love my
<item rdf:about="http://dbpedia.org/resource/Cat">Cat</item>.
(no page there for me!)
This has consequences for textual labels of the URI and I have suggested that there may no longer be real meaning in the term "preferred form" since theoretically, each user could select any label he or she wants.
This is why I think the LC announcement is so important. It must be one of the great repositories of "concepts" in the world. dbpedia is too. Could these be linked? Of course.
I think the power inherent in such a system is clear enough to envision and its possibilities seem endless. To make it complete, i.e. encompassing all pages on the Internet would be useless and unrealistic, but this is where the controlled area (Ross Atkinson’s "control zone") vs. the uncontrolled area would be. I see libraries doing their work in the controlled area, and it seems to me to be the logical extension of librarianship in the digital world.
But...
I may be completely wrong, so if I am, pardons in advance!
Jim Weinheimer
Received on Fri Mar 06 2009 - 06:48:03 EST