|
I know that Tom (Erl) is quite busy with many projects, but I want to facilitate the discussion so that we may go to this issue of how and what.
There is a need to separate how transactions are to occur, and the "why". Let me make this argument in as simplest terms possible.
First we find this separation in the natural world.
This separation maps to one seen in the biology, where function and structure are separated due to an observable stratification in the physical existence, in which function is to be achieved. The stratification theory i propose follows a number of supporting works, including work on self orchestration involved in personal awareness (Hameroff and Penrose).
Second, we do NOT see this separation clearly in the IT markets. Only outside the mainstream is there a stratification separating how and why, as observed in www.secondschool.net.
As I have seen this, cubism as defined by Klausner, is essentially a framework like the semantic framework of Ballard, Sowa, (even Zackman) and also like those based on design principles, such as seen in your recent book on SOA design principles. Unlike software frameworks, like the "Spring Framework" or .NET, there is a clear separation of the how and what in cubism.
So why are there two schools of thought?
In the extensive Coretalk simulation of "How", the mechanisms of "how" is simplified because of the absence of the need for ownership competition over "how".
Imagine a response to a national disaster where the "how" competition never occurs, and the only competition is regarding the "what".
There is a soft competition over the "how" in the virtual nodes of a "pack plate" (Prueitt, 2007) but this competition is driven purely by utility functions defined by the global good and not by private ownership over "hows". This soft competition drives the underlying machinery towards optimality that is neutral to this "how" ownership. Thus "what" ownership can be optimized. As in the (John) Nash theory of economic equilibrium , there is a global good that is accounted for, and this accounting is separated from the private ownership of those means to achieve public good. If the "what" ownership is 100% protected using a instrumented back plate, then we have the Nash model implemented. All that follows this can be discussed using stochastic theory. (Sigh..)
If one is trying to save a population from a storm, the "how" should be optimized.
Patterns are "whats". Design principles are "hows".
The entire subject is of frameworks is seen in the relationship of the chemical table of atoms, the periodic table, and the aggregation of atoms into molecules, as noted by C. S. Peirce.
The structure function relationship is essential to the understanding of organic chemistry.
The structural problem seen in how IT systems now evolve has to do a lot with the fact that "how" is owned not by the agent needing to do, but to the provider of the technology. Simply stated, various providers of technology create separate and non-inter-operative "hows" so that ownership of the "how" drives a market. One sees this manifest in the government procurement practices in the e-Government activities. The how is owned and no one can really get to the what. (We see this also in the ownership of the educational process by institutions that are protecting a failed system of education.)
The alternative ownership pattern is one where the "what" can be tracked so that individuals who have a capability to develop good "whats" can be compensated for this work. This is paradoxical.
The how must be non-own-able, like the air we breath. Only then can the ownership of the "what" create markets based on free enterprise.
The back plate proposal:
Brad Cox's book SuperDistribution applies the concept of a micro-transaction system, a "back plate" (Prueitt, 2007), where are software components are unbundled (separated from group ownership) and provided an internal instrumentation (a code that has to execute in order that the component be used). I long argued with Brad, un-successfully so far, that the concept should be applied to digital product. The digital rights management (DRM) global solution arises if a back plate is given to a marketplace.
The properties of a back plate include
1) a "final and 100% complete" solution to the public ownership of an evolving and contextual substrate, the design elements generated from the application of a philosophical view called cubism. The public ownership becomes "non-owner-able" by individual agreement if the BCNGroup Charter is followed. ( www.bcngroup.org)
2) market transparency over the "whats" within an interaction space, and thus over commodity use...
3) digital property no longer needs to be moved in the internet, because object seeds (generative Encapsulated Digital Properties) can be sent and the object "grown". This reduces the requirements for larger and large information pipes. Bell's inequality in quantum theory is useful in explaining how objects can disappear at one location and re-appear at another location.
All of these features build up a new paradigm in market formation and existence; and breaks down the existing IT service industry market.
We can only hope that something like a federal project creating back plates will allow a demonstration of how these four back plate properties will manifest new less wasteful and more sustainable markets.
On Jan 6, 2008, at 7:17 AM, Sandy Klausner wrote: Tom,
There appears to be many reason's why we should consider some form of collaboration. Please review CoreTalk's web site and I would be glad to review your new manuscript.
Thanks, Sandy
On Jan 5, 2008, at 8:25 PM, Paul Prueitt wrote:
I am reviewing your manuscript, having spent about four hours so far.
The general concern I am having is that reading the book is a long process. There are different audiences, but there is a degree of conceptualization that seems hard to put into my mind all at once. It is clear that you have this all in your mind, and the writing is so excellent that perhaps this concern is not really one that should be taken into account as you complete the work required for publication.
But there may be something to do in the future, when the next administration takes over.
There are two things that I wish could happen, well three.
1) I would like to see all of the mind map like diagrams, example Figure 5.36, converted into web ontology using reification of universals from particular instances.
2) I would like to convert all of the use cases to cubicon expressions.
and
3) I would like to see that part of the development of design principles and design patterns re-expressed with the assumption that a sub-structural component re-use paradigm was always taking care of the issue of re-use. This would follow the methodology illustrated by Klausner's work on CoreTalk (using cubism). There is a vast simplification demonstrated by Klausner's work, at least in principle.
I talk about this issue in my "back plate" design considerations.
I am including Susan Turnbull at GSA and Sandy Klausner and ask that they both send to you addresses so that they may receive a copy of the manuscript, if that is still possible - if there are still copies. I note that both the work by Thomas Erl and Sandy Klausner is quite different from what is covered in the e-government work. I have also long made the argument that presentation and focus in e-government is handled so that alternative to entrenched IT forces can not be seen by government managers.
The back-plate is realized in Klausner's work, so that concept of removing the "how" from "what" in SOA design principles and design patterns is feasible. The "how" is what is being managed by the IT powers, in my opinion. The "what" has to relate to the use cases.
The struggle I am having is that all of Information technology is just to complicated, and in most cases so poorly presented that one can not see the forest for the trees. Your work has always not suffered from poor writing, and perhaps in this way the overly complicated nature of web services, SOA, ontology etc.
The community is learning and things are getting clearer. The issue with respect to how ontologies might be developed from instances is a good example. Classes should not be the consequences of thought processes, without an anchor in instances. The analysis you have done with respect to three case studies (Cutit Saws, Alleywood Lumber, Foresty Regulatory Commission) and their interactions would make an excellent demonstration of SOA use for e-Government.
An exercise to take your three interlinked used cases, and create a ontology mediated demonstration of service oriented computing would be an good exercise for any administration to fund, if there was real interest in making SOA more common.
As you have done, the use cases create adequate illustration of most standardized transactions. However, there is always this question of the use of design principle, design patterns, in cases where the situation involves novelty- such as in crisis management.
***
I do not think that my note is anything but something to think about and perhaps store for future reference.
|