I remember in hey hay days of web 2.0, there was lot of talk
about semantic web. But it seems that just like many other fads, semantic web
fashion (sic!) is also evaporated.
But still, Semantic web makes lot of sense as web is becoming
things oriented from human focused. In
the core of semantic web, lies semantic software.
Before going further, let’s pick some argument which
provides differentiation between syntactic and semantic.
Syntax is what the grammar allows, semantics is what it
means (http://c2.com/cgi/wiki?SyntaxVsSemantics)
But this differentiation does not provide any clue about,
how to make software more semantic. As per
the above statement, every software is semantic.
But, does this full
fill the fully automated internet of things? I do not think so. Then what are
the THINGS/CONCEPTS make a software semantic. In my view software which can evolve
by itself in an ecosystem is truly semantic. Does our computer have reached
that level of maturity where software can evolve by itself? NO.
So, how to make a software self evolve.? I do not know but I
think usage of following concepts in software will push in right direction.
1.
Pattern matching
a.
Text --> Regular expression --> To parse any input text,
stored text
b.
Data àstored,
real time (Complex Event processing)
2.
Pattern discovery
a.
Text (unstructured data)
b.
Data à
stored
3.
XSLT and Schematron
4.
Rule engine
a.
Rules are defined at design time
b.
Rules are defined at run time
5.
Glossary/ontology/taxonomy/data dictionary à not on the piece of
paper but in code. Data dictionary usage can be seen in usage of combo box
instead of free text, option buttons,
6.
Choreography (not orchestration)
7.
Understanding of meta data of various constructs
and giving handle to user to manipulate them to alter behavior of that
construct under the eco-system of given computer ( hardware and/or software)
8.
Artificial Learning (AI) and machine Learning
(ML)
Posts
Semantic Web has to use the existing web data as input for sematics as the existing web is so heavily manifested and has grown to huge proportion that Semantic web may not be able to replace it in entirety. RDF and OWL have jump started Sematic Web, but sematic web's biggest hurdles are 1) Unstructured and Huge exisiting web data (regular links which need to be converted to semantic links 2) Fuzzy States of entities which make semantic meaning difficult
ReplyDeleteDBPedia and SPARQL have been started as Semantic Web Projects.