On 6/24/13 7:00 PM, Dan Brickley wrote:
On 24 June 2013 14:31, Kingsley Idehen <[email protected]> wrote:On 6/24/13 2:14 AM, Michael Brunnbauer wrote:Hello Kingsley Idehen,On Sun, Jun 23, 2013 at 05:32:00PM -0400, Kingsley Idehen wrote:We don't need a central repository of anything. Linked Data is supposed to be about enhancing serendipitous discovery of relevant things.You appear to be arguing against the simple useful practice of communally collecting information.
I am not.
Just because we can scatter information around the Web and subsequently aggregate it, doesn't mean that such fragmentation is always productive.
The simple use of communally collecting information can be varied. To date, only one pattern has been explored with the same results. I am simply suggestion an additional approach. I am never one to propose "silver bullets" since at the core of most of my suggestions lies a preference for multiplicity, flexibility, and choice.
I don't see anyone arguing that the only option is to monolithically centralise everything forever; just that a communal effort on cataloguing things might be worth the time.
It has been worth some time, but it always becomes stale. I am suggesting we add other approaches that haven't been tried to the mix. If everyone simply describes their products, services, and platforms using Linked Data documents we will more than likely realizes that we can dog-food our way to solving an important and thorny problem.
Google already demonstrates some of this, in the most obvious sense via its search engine, and no so obvious via its crawling of Linked Data which then makes its way Google Knowledge Graph and G+ etc..-> http://en.wikipedia.org/wiki/Citation_needed
Google has been crawling DBpedia for years (we do have the logs). I am sure you've seen the technical reports we produce re. DBpedia [1][2]. I am sure you know that cannot be a secret since we do publish HTML docs amongst other formats. It also has Linked Data published via Freebase which I posted a note about re. deconstruction of the obscured Linked Data URI [3].
You've sometimes said that all Web pages are already "Linked Data" with boring link-types. Are you talking about something more RDFish in this case?
Yes. I am saying, let's dog-food i.e., use the technology we are asking the world to adopt etc..
Dan
Links: 1. http://bit.ly/Vie2aB -- DBpedia 3.8 technical report. 2. http://bit.ly/RieuZg -- older report.3. http://bit.ly/LFt9al -- Deconstructing Google Knowledge Graph URIs (note: we'll have an updated unscrambler of these GKG URIs later this week via a new cartridge/wrapper).
-- Regards, Kingsley Idehen Founder & CEO OpenLink Software Company Web: http://www.openlinksw.com Personal Weblog: http://www.openlinksw.com/blog/~kidehen Twitter/Identi.ca handle: @kidehen Google+ Profile: https://plus.google.com/112399767740508618350/about LinkedIn Profile: http://www.linkedin.com/in/kidehen
smime.p7s
Description: S/MIME Cryptographic Signature
