- From: peter williams <home_pw@msn.com>
- Date: Thu, 5 May 2011 22:46:39 -0700
- To: "'Kingsley Idehen'" <kidehen@openlinksw.com>, <public-xg-webid@w3.org>
I can a story here. One of my job missions is to define the utility of cloud storage (for a real estate firm). Wel that sounds easy. Sql in the cloud. Or blobs and tables. Or trivial webservices for CRUD data operations, perhaps with a token in the wwwauth field. POCO entity models, with factory classes that make it all trivial. Throw in a bit of aspect programming, or inversion of control. Throw some more java patterns in, for good measure. But, its not a strategy. Sure, things have moved around, and there is some multi-tenancy. But, no reason to move from my SAN to the cloud. 2 things are starting to change my mind, due to openid and webid. And its not openid protocol nor webid protocol that is the direct change vector. Its what the enable, in their ADVANCED uses. In openid, I was most interested in the XRI world. It was a bit interesting that XRI had what DANE wants to do (a hierarchy of naming records, with signed RRs). But, it was that the openXRI server showed something similar to what I way in data.fm. It was just in "structure" XML as the graphing basis, rather than RDF (as per the data.fm). the openXRI server cleary showed how one made ones own little data cloud, having done a login from an openid. Having played a lot with the more mature products (ODS for the profile manager, and data.fm for the data space), I can see now that semweb is maturing to the point where I was with openXRI. And, it is a bit simpler than the XRI/XDI graphing basis. It's the same, but different. Same enough, so what I HOPED FOR as a side-effect of openid (get to personal named XDI graphs), I can get from the data.fm type world (different, most in the core). SO, Im VERY hopeful. I went through the ODS stuff quite carefully. I got to make 2 certs and 2 keys (bound to different anchored graphs), and Im way past using foaf-a-matic type sites. Its enterprise grade (nearly). I got my data.fm spaces, then, merely by doing webid logon. This is the impressive part. I only hit it today. When I connected the dots that what I lost in XRI/XDI (when openid got reduced to an "irrelevant" SAML-blob-alternative) I got back in data.fm, now Im feeling much better about webid, and its bigger story. Ive got sparql, ive got sparql over https with clients certs, Ive got sparql with a writable port (guarded by webid client auth), and Ive got a XDI-like data space, in which to now build those advanced data-centric apps. And, its all user-centric, in the original sense that identity commons intended for the term (which had nothing to do with PGP saving the Russian refuseniks, etc, by flouting all authorities). I hope some writes a paper on the what it took to realize this topic. It will be worth reading. Now I get to articulate a stragegy for national scale upgrades, and I get to predict when it will occur, and what cost... and guess where to first seek adopting parties -----Original Message----- From: public-xg-webid-request@w3.org [mailto:public-xg-webid-request@w3.org] On Behalf Of Kingsley Idehen Sent: Thursday, May 05, 2011 1:53 PM To: public-xg-webid@w3.org Subject: Re: WebID video and demo On 5/5/11 2:21 PM, Manu Sporny wrote: > The storage mechanism in our particular implementation was chosen just > because it was the simplest way to implement what we wanted. There is > nothing, as far as we can see, that should couple the storage > mechanism to the WebID specification. So yes, we can give people that > choice - but many people are just going to choose the default. This matter is becoming a top priority now that some infrastructure failure points have been addressed. For example, you can use HTML5 WebDB to talk SPASQL to a data space (which can take the form of an RDBMS, Graph DBMS, a file etc..). The channel question will be secured too via WebID. What we'll end up doing is making Flash storage an option :-) Links: 1. http://wikis.openlinksw.com/dataspace/owiki/wiki/UdaWikiWeb/InstallConfigODB CHTML5Bridge - HTML5 WebDB to ODBC Bridges (so SPASQL has a pinhole and more ) -- Regards, Kingsley Idehen President& CEO OpenLink Software Web: http://www.openlinksw.com Weblog: http://www.openlinksw.com/blog/~kidehen Twitter/Identi.ca: kidehen
Received on Friday, 6 May 2011 05:47:08 UTC