[OpenID] W3C TAG recommends against XRI

[Peter Williams]

From: SitG Admin
Sent: Fri 5/23/2008 6:07 PM
To: Gabe Wachob; Dick Hardt
Cc: general at openid.net
Subject: Re: [OpenID] W3C TAG recommends against XRI


As a disclaimer, I do not understand XR*.




Lets try to understand, trying just little harder. I promise -  it will not hurt.

When you were aged about 16, a (hypothetical) school trip probably took you to the British Library (or the Bibliothèque nationale de France, or the Library of Congress  (or... that old Prussian building in East Berlin whose name I've forgotten). If you are Greek, Turk, Persian or from Egypt, you probably went on a symbolic pilgrimage to Alexandria, instead. 

1. In the entrance, you collected a paper map locating various card index systems in rows and rows of old victorian wooden trays (one for physical sciences over there, one for medical science over yonder, art history in the corner, and a * denoting the index for stolen artifacts (pun) from the Παρθενών (pun), located in the basement of the British Library). That map is XRD#1 - a (r)esoure d(descriptor) Your handy pocket OCR scanner converts it into XML. The XML format is traditional and obvious, rather than Semweb's inverted data model.

2. You then went to the rows of "medical reference" card indexes located by the earlier map, and pulled out a card, referencing a shelf location for a particular specialised  "medical dictionary of specialties", one book per town. You chose your home town. That shelving locator card is XRD#2. Your handy pocket OCR scanner converts it into XML. The XRD format is e(x)tensible, so it has no problem adapting to the stuff medical scientists discuss. Its just a bit of XML after all, the latest in a long line of 1970s-era syntaxes for data description.

3. You located the book on the shelf, and looked up address of all the medical folk in the town dealing with pathology X. You OCR'ed the several street addresses, this being XRD#3 file. You even prioritized the mostly equivalent service endpoints/address of the clinics, by language skills. Your handy pocket OCR scanner converted it into XML, of course.

You then sequenced the 3 XRDs files, storing them into 1 memory file persisted to a memory card. Then you posted the collection to yourself, where the fancy name for the envelope is an XRDS.  I wonder what the final S stands for? set, sequence, store... who knows! who cares!

-------------

What is an XRI form of the URI?

The "xri://domain.com/catalog1searchRule*catalog2searchRule*catalog3searchRule" is the roadmap through the 3 card indexes where the * means change catalog. The catalogs obvious depend on the library system you target (@domain.com). Don't take this use of syntax as gospel tho; I probably dropped a *, a @, or an = syntax along the way, somewhere. 

That's is. Difficult isn't it? 

An HXRI is a variant, that uses https. Gets rid of the xri:// bit in front, basically, creating a crutfy URL.

As the above is full of omissions and inaccuracies, I get a C- grade for establishing I can write down "the general idea". If you are currently getting an F grade, a C- is a great step forward. Now read the spec, to get up to a B grade.

Now, if you are building trust models, you should ask yourself: WHICH catalogs are "authoritative" - vs which are a front for the enemy information warfare group deceiving you on good/bad OPs?

You get the A grade when you can properly address this topic of XRI.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openid.net/pipermail/openid-general/attachments/20080524/34ca07cd/attachment-0002.htm>