RE: [xml-dev] Clustering Customization Vs Global Standards

["Bullard, Claude L \(Len\)" <len.bullard@intergraph.com>]

Title: Re: [xml-dev] Clustering Customization Vs Global Standards

The topic gets treated too often as if these concepts are late breaking epiphanies when in fact, they are rehashes of the work done in cybernetics fifty years ago.   Second order systems are known types.  What has changed as with most web technologies is the scale of the possible application given all of the data out there.   Still, Rick is right that getting vetted data is an expensive proposition, much less getting good ontologies and rules.

 

But I find it fascinating that these concepts for self-organizing autonomic systems appear over and over again in different domains.  The act of organizing/identifying clusters by force vectors (strength of relationships) recurs as well.

 

We've thought long and hard about the local functions that enable an XML file to be written and modified.   At the service level, we have the relationships of the societal/business functions that cause these changes to be made.   These relationships also have force vector quantities but I've not seen that discussed much.  One might speculate what real time force vectors are in play in a service-oriented architecture for a situation, that in fact, this is the nature of a situation semantic:  a situation semantic is a high level 'role' for an SOA of a given type.

 

Thoughts?

 

len

---

From: Chiusano Joseph [mailto:chiusano_joseph@bah.com]

<Quote>

the feedback loop involves a lot of information that is not easily or cannot be collected automatically to put in the feedback loop.Sometimes the human effort required to get the information is beyond the resource of the organisation.

</Quote>

 

Yet Bain had no trouble calculating the amount of time they have spent on the problem organization-wide, and sending that information to you:)

 

Joe (sorry, couldn't resist;)

---

From: Rick Marshall [mailto:rjm@zenucom.com]
 

 The big problem with all this stuff (and I've spent a lot of time -
longer than Bain's) trying to figure this one out - the feedback loop
involves a lot of information that is not easily or cannot be collected
automatically to put in the feedback loop. Sometimes the human effort
required to get the information is beyond the resource of the organisation.

I suspect (and it may have to wait for my retirement to find time to
prove it) that a dom aware neural network is the correct answer to these
problems...

Bullard, Claude L (Len) wrote:

> http://hbsworkingknowledge.hbs.edu/item.jhtml?id=5376&t=globalization
> <http://hbsworkingknowledge.hbs.edu/item.jhtml?id=5376&t=globalization>