RE: social recasting and webid - an asde on of RE: Mandatory client supported serializations

ok. ive written something that makes sense to me: http://tinyurl.com/823xzj6.

 

Most of what follows is only partially about webid, but is about getting the baseline (semantic web) out of the way - so I can get back to something I know something about (SSL, certs, key management, critical infrastructure, risk..)

 

Im sorry that it has taken me 2+  years to finally learn to use this tool chain (and build a mental model that works). And, I feel like a dork for having made it appear so difficult. At the same time, you all bear lots of the blame, for giving new terms to old things. The desire to have "wholly" reinvented has not helped this movement. It's clearly recast (not reinvented).


Its all actually very simple - even for some like me who absolutely hated doing traditional SQL (or any other data problem).  And the "foaf project" and even webid now make quite some sense. They are no longer hard.

 

With the semantic web variant of data management problems, since its now just like prolog, I dont mind data, now. I can work with this (and remember, im a data projects total-hater). Thats quite an accomplishment. its made data something even I find interesting, now.

 

In the last 2 weeks, I've got to a decent user-level domination of the profile maker (ODS) and its  various applications some of which manage trust networks (e.g. follower types), and the sparql engine (uriburner), and the enhanced sparql engine (accessible easily at uriburner/isparql) that makes the sparql server act "as a linked data client". (I even understand what one visualizer 9ODE) does, even.) I can now see how the profile maker is distinct from the applications that built upon things like connections and foaf acocunt mapping (from facebook websso), and performing ping back upon configuring "annotations" (e.g. foaf:knows) so bilateral reciprocity might get formed up (to Mo's point), I see finally how its all distinct from a linked data client, too. 

 

These terms were all a mush, before. I cannot make my owl:sameAs annotation actually show up in the foaf card: but it will, given a little magic.

 

but, I get it now. And, I KNOW why it was so hard to take the concepts from the foaf and webid spec and JUST ENDLESS demos, wikis, pages, and writesup ...and translate them into use of tools that do what you expect. There is a LOT of implicit knowhow involved just in using tools (let alone using lower level programming APIs).

 

but, this can be dumbed down, for mass adoption. These tools are indeed what an IT admin needs. And, i CAN imagine ODS being something an realty ASP could procure, probably as a cloud service. It implements something we have wanted to deliver for a while (but didnt know how to find the right technology to deliver the concept). We called it the "agent desktop" I think we found the agent desktop, in ODS. Its more than just hosting a wordpress system on IIS for each realtor, tied to WIF WEBSSO for logon.

 

Finally, 3 years ago we paid a vendor to demonstrate use of another sparql server (and quad store) to intelligently "sponge" live XML data from realty endpoints, which has its own data model imprinted onto XML. This production system delivers billions of bytes a day, in mainframe class syndication processses, to 10,000s of sites.He got quite some way onto inferring the particular tenant's ontology on the fly by walking the schema (and then importing some data). it worked, but failed at engineering. Such was the programmers PURE LOGIC talent that 10,000 nested definitions got formed (mapping a few ENUMs) and consumed all the virtual memory of the machine, when the datalog engine then wanted to process 3 data records). My chier architect laughed, and told me to go away for 3 years, since he had some production SQL stored procedures to get written. Well, that was 3 years ago. Perhaps I can make another attempt.

Received on Monday, 2 January 2012 00:09:43 UTC