RE: Brave new world, OR tagging based on EDIFACT names

[Todd Boyle <tboyle@rosehill.net>]

William J. Kammerer  Friday, July 14, 2000   ebXML Core
> Subject: Re: Brave new world, OR tagging based on EDIFACT names
> 
> Niels Rasmussen suggested that components in ebXML possibly be tagged to
> reflect an EDIFACT heritage (or, more accurately, "baggage").  For
> example, if the data to be expressed as the Estimated date of arrival
> were July 14, 2000, EDIFACT would represent it in the DTM segment as
> DTM+132:20000714:102'.  DTM is the segment tag for Date/Time/Period.

Hello William, I am terrified of all these thousands of tags.  Perhaps
ebXML Core workgroup should consider mechanisms that would accumulate 
statistics of frequency with which data elements are used.

For example, the new user of ebXML might query the polling server, and 
find that the tag "<vendorname>" has been used 50,000 times in the past 
30 days, but the tag "<suppliername>" has been used 80,000 times, and 
conclude that he/she would be more readily understood in ecommerce if he 
used "<suppliername>".  This would reduce the need for experts 
implementing ebXML.  Pretty soon the tag "<vendorname>" would become
extinct.

Obviously, you would collect a couple more facts about each event, 
e.g. the URIs to the schema and definition of the tag.

Any polling mechanism would depend on some convention for voluntary 
reporting by "users" of "content", and some trusted polling station on 
the internet to which their usage "event" could be sent without cost, or 
fear of retribution.  

XML schema for reports or events of usage of content could have wider 
applications in micropayments for music or other copyrighted content. 
Economies may result when many constituencies use the same client and 
server objects to accumulate the event reports. These objects may 
report events in several ways e.g.

 * when content passes thru an abitrary point on the network, or
 * when an application sees a particular content.

Perhaps this idea already exists out there. I don't know.

There is some difficulty today, reaching consensus on basic horizontal
vocabularies for basic data elements, structures such as party,
documents, and 'conversations' necessary to conduct business.  Let
us assume these will be overcome. 

You still need progressively finer attributes for real-world 
marketplaces to function.  One of the problems with B2B marketplaces is 
that products are incredibly diverse and continually changing.  
Customers require to know what they are buying, in granular detail. Who 
is going to be the vocabulary police and naming czars, in every market? 
Do you see why we need a self-healing, self-guided architecture?

Music consumers will love ths.  They can find out what's hot, without
waiting for the music police to "tell them" what's hot.

Distributed file systems such as Freenet and Gnutella suggest a future 
transaction fabric, in which individuals or businesses may execute 
purchases, sales, payments with one another outside the traditional 
webserver or online marketplace. Participants rather than hubs, would 
control who could view their transactions.  Execution and settlement 
would return to the edges of the network. This is geodesic, 
point-to-point commerce.   Let's get on the same page with these 
music people. This is good stuff.

* Todd F. Boyle CPA    http://www.GLDialtone.com/
* tboyle@rosehill.net    Kirkland WA    (425) 827-3107
* XML accounting, web ledgers, BSPs, ASPs, whatever it takes