W3C home > Mailing lists > Public > semantic-web@w3.org > May 2005

Re: a look at SW problems

From: Frank Manola <fmanola@acm.org>
Date: Tue, 17 May 2005 11:56:01 -0400
Message-ID: <428A1411.5010208@acm.org>
To: Linda <linda@alierra.com>
CC: semantic-web@w3.org

Linda wrote:
> When I heard of semantic web for the first time, it reminded me the idea 
> of creating a unique language for the whole world. However, we can see 
> that people still speak English , Spanish, French and so on...  Then I 
> was surfing the web for more information and as I got into the thing, I 
> found the idea really bright, though quite challenging. I believe it has 
> a bright future as it's not something artificial but is a natural 
> extension of the current Web - with better defined information, making 
> it understandable for computers and thus greatly fastening the process 
> of data exchange.

I think you're off to a good start here!

> Universality is the paramount property of the Web, and I guess the 
> essential property of the contemporary world. To succeed, we should make 
> it the basic principle of the research and development in the semantic 
> web sphere. We should unite the best scientists from all over the world, 
> and use the phylosophy from life sciences. For example, most researchers 
> have noticed that the problem reminds of a Godel's theorem in Math: any 
> system that is complex enough to be useful also encompasses unanswerable 
> questions, much like sophisticated versions of the basic paradox "This 
> sentence is false." We know how this has been solved traditionally 
> within quite a narrow approach of each single paradigm. However, with 
> all the knowledge we have gathered all over the centuries, we may 
> overcome this and find new solutions that may be applied universally.

I think we need to be careful not to carry the "universality" idea too 
far, though.  This is not to say we shouldn't be looking to advance 
"universality" (in its various meanings), but sometimes people have the 
idea that the Semantic Web must be "universal" in certain specific 
senses in order to be useful and/or successful (as in your comment above 
about "creating a unique language for the whole world").  I hope no one 
would argue that the current Web is not successful and useful.  And yet 
there is lots of information on the Web that a given person won't 
understand, either because it's not in their native language (or other 
natural language they understand), or because it's nominally in a 
language that they understand, but in a subset that they don't 
understand (e.g., the language of physics, for a non-physicist).  And in 
many cases that person may never need or want to understand that 
information (whether they are right or wrong about that is beside the 
point).  In the same way, the Semantic Web can be hugely useful and 
successful without being "universal" in, say, the sense that a given 
piece of software can "understand" (in a meaningful sense of software 
understanding anything) all the information that is "relevant" (by human 
criteria) to what it is doing that may exist on the Semantic Web.  Mind 
you, it would be nice if it could, and obviously we should work in that 
direction, but I don't see this as a necessary success factor for the 
Semantic Web, any more than the corresponding capability is for the 
current Web.

> Among the 3 basic components of the semantic web, personally, I see most 
> difficulties in ontologies. There have always been problems with 
> defining relationships, sometimes the extentions is different in 
> different dictionaries. I didn't have a chance to read much about that 
> and honestly I prefer communicate with people instead of looking for 
> specific information in a book or serching the web. How can we overcome 
> the difference of extentions? What are we going to do with polysemous 
> words? If anybody can give me some information or a link to a 
> related article, I will appreciate it a lot.

There's a lot of literature in related areas.  Plugging search terms 
such as "ontology integration", "database integration", or 
"heterogeneous databases" into Google will find a lot of relevant 
material (as well as the usual irrelevant stuff!).  Certainly the 
database community has been looking at some of these issues for a long time.

Received on Tuesday, 17 May 2005 15:56:17 UTC

This archive was generated by hypermail 2.4.0 : Tuesday, 5 July 2022 08:44:53 UTC