The Semantic Web: Everything you wanted to know but were too afraid to ask.

by Dan Zambonini

There are numerous misconceptions about the Semantic Web, largely caused by a misunderstanding of its aims and technologies. I've created this simple FAQ help dispel some of the myths.

7 Comments

Taylor
2006-12-18 13:11:16
"There are two basic approaches for making this data available to the Semantic Web."


#3, put it directly in the webpage. Regarding RSS, isn't that a bad sign? It ~WAS~ RDF. That one still leaves me disappointed and a bit hesitant to believe RDF is the best way to apply semantics to the web.

tndal
2006-12-18 14:43:37
The Semantic Web (SW) never worked, won't ever work and was designed ignoring basic principles that prevent it from ever working. There are good reasons why Google's people chuckle at the SW and chide Tim Berners-Lee (TBL).


SW papers over all previous research (by better people) in ontologies, language, logic and AI. Somehow, if we all put our data out there, we are told that all disparities in different ontologies (definitions of terms) will somehow disappear and the SW will flourish. Of course the SW people will rush in and say, "No, No, that's not what we mean!" but they cannot provide a logical, meaningful description of what the SW should do that that truly makes sense. It's all vague puppies and butterflies.


The SW is TBL's vanity project and we should treat it as such. TBL is a nice guy but not nearly bright enough and far too stubborn to admit that, while he got the browser right, he got what he calls the SW all wrong. SW is TBL's bong pipe dream and he's never been able to admit that his idea was stillborne. His message is constant: inhale the smoke and drink the Kool-Aid.

stelt
2006-12-18 15:53:17
I wrote this some time ago:
===========================


An attempt at a very short explanation of rdf, the important ingredient of the semantic web:


- triple(s): (instance A_thing [of type A_class]) {HAS (some sort of relation Semantic_link) WITH (instance B_thing [of type B_class])} {}:1 or more of these
- Semantic_link is not a hyperlink in the traditional sense, but it is defined by/at a URL and is directional
- Semantic_link, and if used (to make things explicit in the data, not implicitly in the code using the data) A_class and B_class, are to be defined in an ontology, aka RDF Vocabulary, aka RDF Schema
- An ontology is the definition of a group of classes and their relationships (usually for just a certain domain or field of expertise, FOAF being a simple example)
- This definition can be written in a 'standard' way by using the OWL notation.
- When different knowledge-/databases have ontologies that intersect or are the same, you can combine the data by finding A_thing or B_thing mentioned more than once


One could make all relations in a SQLdatabase explicit by transforming it into an RDF database when you define the relations between the columns and use that in the transformation script.
With Prolog or Sparql taking advantage of the fact that the first-degree relations are already in the data, this allows the programmer to program in a syntax/language that is closer to what (s)he's trying to model.


2006-12-19 04:32:47
> you'll still get all the benefits of XML


Can I still use XSLT?
Can I still use XML Schema?
Can I still use CSS?
...

Charles Young
2007-01-03 10:04:17
Is Microsoft really investing seriously in RDF? Maybe, but I don't know what the evidence is. The Profile Manager you mention is part of their little-known Connected Services Framework which, to date, has been mainly targeted at Telcos and used for implementing provisioning systems. Quite how CSF fits in with Microsoft's 'connected systems' strategy is a matter of some debate. I'm fairly sure that CSF isn't seen as 'mainstream' within the MS Connected Systems Division. I don't know of any other Microsoft initiatives that are using RDF, though a quick Google search suggested they may be using it somewhere inside Vista. Microsoft is not monolithic in terms of its product groups, and the fact that one group is supporting RDF and SPARQL for a straightforward profile database does not suggest a corporate commitment to RDF from Microsoft as a whole. I wouldn't get your hopes up to high on this one. Microsoft is a large company, and it would be statistically highly improbable that there aren't some RDF/SW enthusiasts within its ranks. It will also employ RDF/SW sceptics (of whom there are many). Like any other software company, I wouldn't expect MS to seriously invest in these specifications until it can see a clear business case and emerging market for the RDF-based SW.
mawar
2007-03-29 22:32:40
thanks a lot for your opinion and explaination abaut semantig but why don't you try to make and explanation in edducation side another student able to understand too..
gongrats and keep on!!
jenny
2007-06-15 16:29:21
Dan your articles are always excellent reading. Person can learn so much just reading it. I hope that in the future your text will be just as good.
jenny