Kingsley, I like this line of argument. When I try to imagine myself
back in 1995 discovering ODBC (versus 4GL's against specific, if
SQLable, databases), I think I could have grokked a lot of the Linked
Data arguments from this comparison.
I like specifically this 'superkeys' bit.
What I find is a little lacking, though, is the idea of a lot of those
keys already existing (in open datasets). I don't think I would get that
from:
"Web data (e.g., data from DBpedia) added to the mix" and
"using the Virtuoso ODBC driver against public services such as URIBurner"
I'd like to see something like "saves you from having to re-invent keys".
Just a quick first impression.
Barry
On 30/03/2012 21:47, Kingsley Idehen wrote:
All,
I've written a detailed post [1] about the timelessly important issue
of Open Data Access as it relates to Linked Data. I've responded to
snippets from a video [2] produced by one of our data access
middleware competitors turned co-opetitor [3].
Linked Data is best sold to the enterprise as an improvement over
ODBC. Basically, an improvement that takes them from Open Database
Connectivity to Open Data Connectivity, courtesy of Web ubiquity.
To the enterprise customer, an HTTP URI is a powerful Data Source Name
(DSN) mechanism. Its also a Super Key delivery mechanism since all
keys actually resolve to records that are intensional (description
oriented propositions in 3-tuple form).
What I outline above is a "deceptively simple" value proposition that
any enterprise customer will comprehend, naturally.
Links:
1. http://goo.gl/1dO33 -- Is Open Database Connectivity (ODBC) Dead?
2. http://youtu.be/kxjxlo5yTVk -- Video in full .
3. http://dbpedia.org/page/Coopetition -- Coopetition explained .
