Re: Alternative Linked Data principles

Hi,
Kingsley - Your response was clear, concise & well put -- Thanks.  For people familiar with RDBMS wanting to better understand what makes a linked data approach useful, that was a great comparison. 

Luca -  Sure, data has a wide range of mechanisms in which it can be modeled and presented. And yes, many would agree -- it is important to view the world from different perspectives.  That said, the people on this list, many of whom have done advanced research in information retrieval, library science, AI, computational linguistics, etc. would probably also agree that the ship has sailed for HTTP URIs -- the mechanism has proven to be the most flexible way to express a global, interlinked database.

Rather than asking whether "UDP works better for achieving federated queries", perhaps the question should be, 'how can I take the most scalable information system known to humankind and extend it for something useful (in whatever field you are working in.)' 

IMO, it isn't a protocols question any longer -- rather, it is a question of leveraging what a bunch of smart people before you have done & making it even better & more useful to share knowledge.  I'm a fan of extending rather than recreating the wheel, over & over again.

Cheers,

Bernadette Hyland
CEO, 3 Round Stones, Inc.

http://3roundstones.com
http://about.me/bernadettehyland 

On Apr 28, 2014, at 12:54 PM, Kingsley Idehen <kidehen@openlinksw.com> wrote:

> On 4/28/14 11:55 AM, Luca Matteis wrote:
>> Thanks John but not really.
>> 
>> I was specifically looking for research that wasn't based on protocols
>> such as HTTP, URIs and RDF. But that is still in the field of
>> achieving a global interconnected database.
> 
> I'll try:
> 
> de-referencing an identifier en route to a description (some system specific representation of a data structure) at an address dates back to the advent of computing.
> 
> For instance SQL RDBMS engines do this using keys (Primary, Foreign etc..). The problem is that these identifiers are literals and de-reference is an act of DBMS engine specific name->address indirection. Which is why ever SQL RDBMS that only implements literals based identifiers or DBMS specific reference types (as per ORDBMS products of yore)  is a data silo vector ++
> 
> It's also why all applications developed using existing native (relative to host OS and runtime combination) programming language are also data silo vectors, since they also use local (or app specific) mechanisms for name->address indirection and eventual structured data representation, following de-reference.
> 
> In a sense, what John is saying to you is as follows:
> 
> HTTP URIs are identifiers that make all of the above webby i.e., no DBMS specific name->address based indirection, no operating system + programming language specific name->address based indirection, it all just happens via HTTP URI based hyperlinks which are based on open standards (i.e., platform agnostic).
> 
> The virtue described above is the kernel of the Web architecture. That's what makes it tick. That's why we have a World Wide Web that's dexterous enough to handle past, present, and future data access and integration challenges without introducing platform specific data silos.
> 
> The novelty of HTTP URIs lies in their truly unique ability to enable mass data de-silo-fication, at Web-scale.
> 
> I hope this helps.
> 
> 
> [1] https://www.youtube.com/watch?v=i49_SNt4yfk -- Binky on Pointers (names or "*") and Pointees (addresses "&")
> [2] http://bit.ly/1rk0s76 -- Recent post about Oracle data-de-silo-fication via HTTP URIs  (note: Green Links [local] vs Blue Links [global] )
> 
> Kingsley
> 
>> 
>> I know webby standards are implemented so no need to reinvent the
>> wheel, but I think it's healthy to look things from a different
>> prospective; who knows maybe UDP works better for achieving federated
>> queries. Or maybe triples aren't really the only way to represent the
>> real world.
>> 
>> Luca
>> 
>> On Mon, Apr 28, 2014 at 5:41 PM, John Erickson <olyerickson@gmail.com> wrote:
>>> Luca, I think you are not asking quite the right question; I think
>>> what you want to ask is whether the Linked Data Principles can be
>>> applied to different...
>>> 
>>> * entity identifiers...
>>> * protocols with which to resolve and retrieve information about those entities
>>> * protocols with which to retrieve manifestations of resources
>>> associated with those named entities...
>>> * file formats with which to serialize manifestations of resources...
>>> * standards for modelling relationships between entities...
>>> 
>>> The value of the Linked Data Principles as bound to "Webby" standards
>>> is that they are specific and readily implemented; no make believe...
>>> 
>>> John
>>> 
>>> On Mon, Apr 28, 2014 at 11:23 AM, Luca Matteis <lmatteis@gmail.com> wrote:
>>>> The current Linked Data principles rely on specific standards and
>>>> protocols such as HTTP, URIs and RDF/SPARQL. Because I think it's
>>>> healthy to look at things from a different prospective, I was
>>>> wondering whether the same idea of a global interlinked database (LOD
>>>> cloud) was portrayed using other principles, perhaps based on
>>>> different protocols and mechanisms.
>>>> 
>>>> Thanks,
>>>> Luca
>>>> 
>>> 
>>> 
>>> --
>>> John S. Erickson, Ph.D.
>>> Deputy Director, Web Science Research Center
>>> Tetherless World Constellation (RPI)
>>> <http://tw.rpi.edu> <olyerickson@gmail.com>
>>> Twitter & Skype: olyerickson
>> 
> 
> 
> -- 
> 
> Regards,
> 
> Kingsley Idehen	
> Founder & CEO
> OpenLink Software
> Company Web: http://www.openlinksw.com
> Personal Weblog: http://www.openlinksw.com/blog/~kidehen
> Twitter Profile: https://twitter.com/kidehen
> Google+ Profile: https://plus.google.com/+KingsleyIdehen/about
> LinkedIn Profile: http://www.linkedin.com/in/kidehen