Monday, 20. October 2008

A problem of semantic web: not providing XML value

Today in the morning, I had a sudden "insight" about one of the problems of RDF and Semantic Web: it misses some of the value that XML offers. This is what my daily commuting bike ride is for, thinking...

The adoption of Semantic Web rises and falls with the adoption of it in standardization bodies. For example the Oil&Gas industry of Norway is thinking about Semantic Web, and I have recently been talking with people from the automotive supplier industry about Semantic Web. To interchange data in a business-to-business environment you would expect that RDF has more features than XML, but in fact, it doesnt.
  • RDF is less expressive than XML. One example: you can't define pattern in RDF. Look in the XML spec, there is much more of it missing in RDF.
  • RDF is not validated. Although in theory, it is possible to validate a file for semantic correctness, nobody does that because of the open world assumption. Hence, there is no validation of the XML in mainstream applications.
So, if you are an industry, you already havean XML based standard, moving to RDF without the expressiveness of XML and without the notion of validation is tricky. RDF should have more features, not less.

Also the stack of XML technologies must be embraced better, for example a XSLT-friendly RDF/XML serialization. Please, dear reader, solve these problems and make a company around it.
QR barcode by i-nigma.com/CreateBarcodes
Cameron Hunt (guest) - 22. Oct, 18:00

Overloading Structure

I would argue that some of the complexity of XML (and XSD) is an attempt to overload what is essentially a standard for data structure to include components that represent data models.

When RDF is combined with RDFS and/or OWL, there is a greater ability to perform semantic validation while allowing for (what I see in my environment) the constant changes and additions to upstream data sources. While open world assumptions can make it difficult to detect data out of place, it can provide an easier framework for data models that are constantly changing (provided the downstream consumers have a plan in place for how they handle new data).

leobard - 24. Oct, 17:12

the argument is theoretic

The critique is: nobody validates RDF and there is no pressure nor tools to validate.

In theory... everything is possible, of course, but that does not help anyone buying a RDF validator or stopping people from writing
<//www.example.com> rdf:type "muahahaha".
icon

semantic weltbild 2.0

Building the Semantic Web is easier together

and then...

foaf explorer
foaf

Geo Visitors Map
I am a hard bloggin' scientist. Read the Manifesto.
www.flickr.com
lebard's photos More of lebard's photos
Skype Me™!

Search

 

Users Status

You are not logged in.

I support

Wikipedia Affiliate Button

Archive

October 2008
Sun
Mon
Tue
Wed
Thu
Fri
Sat
 
 
 
 1 
 2 
 3 
 4 
 5 
 6 
 7 
 8 
 9 
10
11
13
14
15
16
17
18
19
21
23
24
25
26
27
28
29
30
 
 

Credits


austriaca
Chucknorrism
digitalcouch
gnowsis
Jesus
NeueHeimat
route planning
SemWeb
travel
zoot
Profil
Logout
Subscribe Weblog