OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help



   RE: [xml-dev] Re: Major Historical SOA Milestone Today

[ Lists Home | Date Index | Thread Index ]
  • To: <andrzej@chaeron.com>,<xml-dev@lists.xml.org>
  • Subject: RE: [xml-dev] Re: Major Historical SOA Milestone Today
  • From: "Bullard, Claude L \(Len\)" <len.bullard@intergraph.com>
  • Date: Fri, 12 May 2006 14:40:08 -0500
  • Thread-index: AcZ1Df3UIIAMClQ/SJSQ9s/EOPMpfQA5I0YA
  • Thread-topic: [xml-dev] Re: Major Historical SOA Milestone Today

This strays away from the topic a lot.  Delete if you are busy.

From: Andrzej Jan Taramina [mailto:andrzej@chaeron.com] 

>>The quote you make is the defense of irrational behavior.  

>No, it is not.  The quote I made is a recognition of human falibility
>reality.  Many technocrats build wonderful monuments to IT
>goodness, and the project fails because they don't take into account
>messy world of real humans.

Life among the mammals.  I fully understand that.  Sterling is wrestling
with that today.  Messy situations emerge and we respond with technology
that might fix the problem or the cure is worse than the disease.  Yet
that is not just a problem of technology but of law, institutions,
electors and the whole swath of human-driven systems development.  OTOH,
as soon as I believe perception is reality I mount the slippery slope
where propaganda and garbage-in-garbage-out rule my actions.  Can you
say, "Squidgy"?  Yes, these look like American politics but the Brits
taught us how during WWII and Britain is a more wall-to-wall wired
country than the US.  The Americans are just now waking up to a trend
that has been grinding ever onward since WWII in this country and WWI in
Britain and possibly before elsewhere.

Who makes it possible?  We do.  The members of this list particularly.
We do.  Why?  That is where perceptions enter the picture.  If one wires
up a world wall-to-wall for convenience and access to ever greater
groups and resources, one also makes all the bad sides of that possible.
We want the credit for the web but not the blame.  OTOH, should we
exhume Alexander Graham Bell and burn his bones?  Shall we burn Cerf and
Berners-Lee in effigy?  No gain there. We are all forging our own chains
for the convenience of a public that is only now waking up to the facts
that massive communications make massive data collection possible.   If
you don't want that, throw away the pod on your ear and start paying
cash for all debts.

It comes down to individuals and administrations because it is not what
a technology does but the way that we use it that matters.  The NSA
collects the data, and yes, the way packets work they have to collect
all or most of it to do their appointed jobs.  What matters is what
questions are asked of the data that they are obligated to answer.  For
that, look to the folks making the requests.  To Sterling all I can say
is, "Yes it sucks.  Be careful who you vote for.  Pay attention.  Abuse
is not a technology but a policy."

>IMO, the concepts of SOA are not domain specific, contrary to your
>since it is an IT-level architectural concept and set of principles.  

That is one viewpoint.  From another it is a means to specify a business
type and procure based on it.  No CompSci guys need apply because they
aren't comfortable with that.  Fortunately, if the concepts are coherent
at every scale of the viewpoint, they don't have to.

>So we should be able to agree on some common definition of SOA.  
>Unfortunately the politics (another real world messy issue that is 
>inescapable) of the different standards bodies and 800-lb vendors are 
>typically in conflict and thus I doubt we'll see a common definition in
>short term.

A common agreement is like a standard:  it is the codification of a Nash
equilibrium.  When the agreement is to everyone's advantage and there
are no reasonable advantages to change, it will be agreed on.  Now, how
long that lasts is anyone's guess.   I believe this kind of agreement is
like a DTD or Schema:  it should be simple, it should be specific and it
should get rid of any extras that meet rare exceptions.  See Rick
Jeliffe's excellent article on xml.com for that point of view.

>With luck, maybe the "best" definition will win in the marketplace of
>before organizations get fed up and label "SOA" as yet another failed
>"magic bullet".


>Feelin' lucky? ;-)

Learned to make my own luck by knowing which viewpoint is appropriate
for a transaction.  I understand SOA from several points of view and am
comfortable with the definitions.  Given some contract, I know how to
spot a mixed viewpoint and how to detect leakage.

> Keep in mind, if we apply the rule of the wisdom of crowds,
> we accept the existence of God, UFOs, ghosts, and that Chris deserved
> be kicked off American Idol.  From the sublime to the silly, crowds
> aren't that discerning.

>True enough...and your point is?  

Read the fine print.

>Whether a crowd (how about we call it a "market") is wise or not is 
>irrelevant if they label your "product" (ie. "SOA" in this case) hype
>avoid it like the plague.  You can rail against the market's lack of
>all you want, while your "product" languishes and dies on the vine.

True for some n of 'they'.  That is a FUD tactic.  Microsoft used it and
the open sourcers do the same.  "If you ain't like me, you may not be at
all."  The real question is what would it be like if we didn't have SOA
concepts?  Possibly not much different because the concepts bubble up
out of the reality of OOP-over-DB and it is better if the viewpoints are
coherent.  Everyone wins.   On the other hand, SOA is the market winner
these days.  Explaining REST takes too much knowledge in the audience.
So the real battle there is not in the market but in the
implementations.  I repeat, we've had this discussion and it comes down
in most cases to how much a developer wants to roll their own smokes.

>> 2) "For every complex problem there is a set of simple solutions and
>> they are the only things that provably work."

>Companies rarely do anything from a sense of honour, including
>Another one of those "real world messy" things.

Individuals and administrations:  justice is not found in nature.  A
company is usually not complete in those respects but the majority are
not evil.  The law of averages catches up to them.  Nor do any companies
stay pure.  See Google.  So pick your partners and the situations well.

>BTW...I subscribe to the 2nd approach for the most part.

Me too.

>> Joe is trying to clear it up.  It is a good idea.

>On that we agree wholeheartedly.

Individuals make a difference.  When they can also tap the list for
spirited debate and ideas, the chorus can help out.  That's what we do
here because XML is pretty much done.




News | XML in Industry | Calendar | XML Registry
Marketplace | Resources | MyXML.org | Sponsors | Privacy Statement

Copyright 2001 XML.org. This site is hosted by OASIS