Cory's excellent note is copied below (for my bcc) (01)
My note will be brief. When a group of four or five of us really work hard
to get to root causes we can sometimes form a new community of practice.
This is what we are looking for, something really new. (02)
Now for some "corrective action". (smiles) (03)
Cory, in a different note you were critical of John Sowa's questioning the
motivation of those who developed the current W3C RDF/OWL standards. I know
that John was not intending to question motivation, and I remember well the
interactions last year I had with Brand and Patrick in the ONTAC forum.
Questioning motivation might be truthful, but it is often not a community
building exercise. John pointed this out to me, last year. (04)
What John is doing is describing history. The fact, as you state, that
people work hard and are committed to their sense of legitimacy is something
that can be observed. But like the algebra student who "tries hard" but
cannot produce the answer, the W3C is building on precisely what John has
described as a matter of history. (05)
Do we as a society merely accept arithmetic illiteracy on the part of almost
all citizens, or do we question the educational process? I agree that John
did not, and does not now have something that is an alternative. Common
logic is not an alternative to RDF. But I do, and Azamat does, and five or
six others do. These are true fully blown alternatives... that are (each)
ignored by funding mechanisms... up to now. Will this change? I believe
so, and soon. But I may be wrong. (06)
I might say that "we" work just are hard as those who defined the W3C
RDF/OWL specifications. (07)
The why , no alternatives are funded, is interesting and I have written
extensively about this topic.. and will not repeat any of that here. But
"motivation" is a large part of my answer. (stated bluntly.) (08)
The paper by Leonid (09)
http://ototsky.mgn.ru/it/21abreast.htm (010)
is describing a history that has no. or little critic of bind alleys or
incorrect development pathways. (If I read his work correctly... perhaps
you will correct me, Leonid). Leo O (at Mitre) does the same thing, as does
Jim H. at UM. These are histories that have little or no insight into
possible wrong paths and wrong development processes. Everything done is
good, and more good is only done by following the current pathway. (011)
There is never a paper with the title "How we spent 30M on something that
sits in a desk drawn unread; and got away with it (again)." Or "why the
x-y-z model we invented did not work". (012)
This telling of positive histories is very IT profession centric, and is in
some people's mind not helpful to an industry whose back is against a wall
(even if some do not really realize it). We (the IT community) must raise
our eyes to the community of human being whose lives are degraded by an
absence of real information science. (One looks positively to library
science for help..... any one here in that field?) (013)
For example, the take home message I get from Leonid's paper is that one has
to wait at least 30 years with no expectation that a by-pass to entrenched
conceptual and technical problems will shock the market. (many of us do
not think humanity will survive the next 30 years without a deep by-pass of
the current IT mess.) (014)
Not being able to seriously question the process (which is what John is
doing) is not helpful if one is thinking outside of the IT community. (Not
questioning may be helpful to the IT sector. Since it is clear that if
information science is not corrected that the over all (short term)
consequence is beneficial to the IT industry - since building information
systems that fail will be accommodated as "normal and expected".) More work
to go around. (015)
The two categories of community response to challenge are incomplete. But
war is not the only other option. (016)
You state below: (017)
" However, when a member of the community violates the understood fairness
(legitimacy) of the community, one of two things happen. One outcome is
that it forever disrupts the community, and the community eventually
ceases to exist. There are a few examples of this in [1]. (018)
The other example of this outcome is that the community sees a need to
more clearly state these shared concepts of legitimacy. " (019)
In the case of IT, this community may need to be ignored, and allowed to
dissipates eventually not because the community sees a disturbance but
because there are other more interesting things to do. This dissipation may
take 30 years. But the break through is near by. (One year, six months..) (020)
Thus I talk about the development of a K-12 curriculum in the knowledge
sciences as part of the proposed national project to establish the knowledge
sciences. The curriculum development will focus on things like
non-locality of meaning and pragmatics as discussed in the BCM
specification. (021)
Are you willing, Cory, to leave some of the difficulties of legacy behind
and move with us into a brave new world? (022)
-----Original Message-----
From: soa-forum-bounces@xxxxxxxxxxxxxx
[mailto:soa-forum-bounces@xxxxxxxxxxxxxx] On Behalf Of Andrew S. Townley
Sent: Thursday, March 30, 2006 6:48 AM
To: Service-Oriented Architecture CoP
Subject: RE: [soa-forum] Business Need for SOA (Was SOA Semantic Variation ) (023)
Ok, some sleep helped a little. I think the answer is to figure out
what SOA *really* is (as I believe was suggested eralier). (024)
Advance warning: I'm not really going to be talking about technology at
all, and I'm going to relate a few things which some may find a bit
tangential. However, I think if we can solve this, the rest is going to
be a lot easier. (025)
In 1722 Benjamin Franklin wrote (as Silence Dogood), "'I have...found by
Experience, that it is impossible to talk distinctly without defining
the Words of which we make use'". (026)
So, that's the challenge. I think what we've seen thus far in the
discussion has really been somewhat affected by this issue. We're
making progress however. :) (027)
I think the most important term we've been using within this context is
the word 'community'. However, I think how it is understood (or not)
has massive implications about how successful our efforts will be with
SOA (industry-wide). For many years, I have had the intuitive feeling
that you can't legislate morality--it comes from something much more
fundamental than laws. Recently, from discussions on the XML-DEV
mailing list, I was directed to the writings of Aldo de Moor. In them,
I finally realized where my intuition originated: the concept of
legitimacy. (028)
Whitworth and de Moor define legitimacy as the common sense of fairness
which binds a community together [1]. As such, it is what allows the
community to grow, because continued acts in and towards the community
based on the same sense of fairness end up building trust amongst the
community. The more trust, the more openness, communication and
cooperation, and therefore the more benefit to the community as a
whole. Those of you familiar with Systems Thinking will recognize this
as a reinforcing system. (029)
However, when a member of the community violates the understood fairness
(legitimacy) of the community, one of two things happen. One outcome is
that it forever disrupts the community, and the community eventually
ceases to exist. There are a few examples of this in [1]. (030)
The other example of this outcome is that the community sees a need to
more clearly state these shared concepts of legitimacy. In society,
this is done by passing laws. In computer systems, it is done using
access control and by introducing other design constraints. (031)
In society in the large, option one really isn't a runner, so more often
than not, the second one is chosen. I suppose, in the interests of
completeness, I should point out that there's a third option, war, but I
don't think it fits into the current discussion. (032)
Assuming that the second option: making laws to enforce the understood
legitimacy traits of society, is chosen, this eventually leads to an
evolved framework defining various aspects of legitimate behavior within
the community. It works, because they are, by definition, *legitimate*,
i.e. fair. People who understand legitimacy in a community normally
don't have to be told what is right and what is wrong for every decision
they make. When implemented properly, everyone's actions generally
reinforce the legitimacy of the community, thus contributing to the
reinforcing system described above. I have written about this
previously in the context of enterprise security [2], but not in this
pure of a form. (033)
Of course, there are always participants in a community who don't
subscribe to the same legitimacy principles. That's where the laws kick
in and either prevent them from doing any damage, or reactively attempt
to prevent others from following a bad example. Still, if coercion is
the only option, this approach will eventually break down as described
in [2] due to the trait of tolerance. (034)
When the system is working correctly, very little interaction with the
codified principles of legitimacy within the community (laws) is
necessary by the participants. I'm sure not many of us could, from
memory, list the exact details of the laws that exist in a locality, let
alone on a larger scale such as a state or a nation, but that fact does
not generally prevent us from going about our daily business. Thus, I
posit that this is primarily because we subscribe to the legitimacy
principles of the communities in which we participate. (035)
If this is accepted, then what are the outcomes of the community? The
main one is that the members of the community are fairly autonomous.
Either implicitly or explicitly, they understand the context in which
they operate, altering their behavior to conform to the legitimacy
principles required for the various communities in which they
participate. Before long, this alteration happens automatically,
without conscious thought from the participants. Do you talk to your
mother the same way you talk to your friends? (I do, because I have a
cool mom, but most people probably don't. This behavior is an example
of what I'm talking about). (036)
One of the largest benefits of autonomy is the ability to work
co-operatively to accomplish a larger goal. However, this "directed" or
"cooperative autonomy" is only effective if everyone fully understands
what they are supposed to do and have the ability to do it. (037)
Bringing all this back a little closer to the topic at hand, the degree
of autonomy in a system is directly proportional to the scalability of
that system. I have elaborated on this concept from a messaging point
of view [3]. However, what struck me this morning when I was thinking
about this conversation was this relationship: (038)
scalability => autonomy => legitimacy (039)
Or, that in a system, scalability implies a certain amount of autonomy,
but that this autonomy is only possible if the autonomous entity has an
awareness of and subscription to the legitimacy principles of the
community in which they are participating. (040)
Bringing this back even closer to the current discussion, my use of the
road metaphor in the earlier mail was a better of example of this
thought than I realized. I said: (041)
> > The SOA part kicks in when I can drive from McD's, to BK, to Wendy's to
> > Arby's, or wherever, and send the same message and expect the same
> > response. In W3C-speak, ordering food is the abstract service and each
> > of the fast food places is a concrete implementation of that service (an
> > agent). The SOA tells me how I can get from point to point, what rules
> > I must follow, and maybe even gives me directions. It may also allow me
> > to get fuel, buy a candy bar or even a newspaper--if those messages
> > existed and there was an agent implementing a service providing them. (042)
Relating the architecture of the SOA back to the society in which we
live and adding in the concept of legitimacy, the traffic laws say where
and when we can drive from point to point. Mostly, people obey these
laws, but I would say that the most fundamental ones (speed limits don't
count here :D) are "obeyed" less because they are laws, but more because
they resonate with our innate sense of fairness. (043)
So, these are the reasons that I actually don't agree with you that
requestor/provider roles/concepts are too low-level within the context
of the SOA. To me, these provide the power of the SOA. This leads me
to conclude that the architectural constraints of the SOA define the
legitimacy principles of the environment, in a similar manner to the way
Whitworth and de Moor say that the programmer is ultimately defining the
legitimacy principles of the social interaction systems they design and
build [1]. (044)
If a framework for what is and is not legitimate within the SOA
community is defined, the participants within that community are free to
be autonomous within that community as long as they follow the rules.
This autonomy within the community allows the community to scale in the
same manner as the Internet [3]. (045)
Note as well that these rules of legitimacy are not restricted to the
software agents acting within the community. It also applies to the
manner in which the human participants interact with each other and with
the software agents as part of the community. This is a model that has
been proven to work (to various degrees) in both computer systems (the
Internet) and society. Managing autonomy is harder than some of the
alternatives, however it is the only way to provide sufficient
scalability within the community to accomplish anything meaningful. The
balancing system at work here is dealing with those that don't conform
to the legitimacy principles. Once these cases are identified, steps
are taken to augment the framework, and life goes on. (046)
This is exactly what happens between government and the governed. If
one sees the other violating the community's legitimacy principles past
a tolerance point, action is taken--in either direction--to re-balance
the environment. I believe that this is the model that SOA must follow
if it is to be successful. (047)
In large deployments with hundreds of services, thousands of people and
millions of messages, it has to have the ability to adapt on its own
behalf--within the framework that defines the architecture of the
community. When this behavior does not occur, the scalability, and
therefore the effectiveness of the architectural style, becomes
constrained. If that happens, we'll be here in another 5-10 years
arguing about the "next big thing" vs. "the last big thing" because we
will have reached a point where the SOA model is no longer useful. (048)
As the Internet has proven, I think it is possible for this type of
community architecture to exist. We just can't repeat some of the
mistakes of the Web which did not consider the legitimacy concerns of
the participants and which has led to some of the issues we have today's
society (e.g. widespread identity theft and fraud). (049)
By allowing the business processes and roles (as you use the term) to
fully participate in this community *as peers*, I believe that SOA is an
architectural style which can achieve the business needs of
collaboration, integration and agility. We just have to create the
right community. (050)
ast (051)
[1] http://www.starlab.vub.ac.be/staff/ademoor/papers/bit03.pdf
[2] http://atownley.org/published/2005/06/ISB1005AST.pdf
[3]
http://atownley.org/2006/03/self-directed-messages-the-key-to-scalable-soa/ (052)
_________________________________________________________________
Subscribe/Unsubscribe/Config: http://colab.cim3.net/mailman/listinfo/soa-forum/
Shared Files: http://colab.cim3.net/file/work/soa/
Community Portal: http://colab.cim3.net/
Community Wiki: http://colab.cim3.net/cgi-bin/wiki.pl?AnnouncementofSOACoP (053)
|