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] Simplification and Composability (RE: [xml-dev] SOAP Messa

[ Lists Home | Date Index | Thread Index ]

So he discovered Cox and sex about 12 years after 
the last generation did.  It will be fun to see 
what analogies he draws.

I used Brad Cox's book extensively when 
I was writing Beyond The Book Metaphor and the Enterprise 
Engineering papers.  The next step was ecological metaphors 
because they enable the concepts of energy budgets.  Software 
systems barely need that but because software system 
implementations always have social components, it is a 
useful analogy.   SoftwareICs tripped over the lack 
of common object models (what Java and C# are there 
to cure but...).  It's no good to have a softwareIC 
if the board one mounts it on isn't fit or it is 
too densely populated.
 
A parameter list is just a microdocument where 
the only consumer is software.

What is more interesting 
is the process of getting agreements that the 
thing named by the URI is the same thing for 
all parties.  All local.  It is easy to designate 
an interface.  It is expensive to maintain it 
the more specific and less applicable it becomes 
but it is better for a shared real time system 
if it is small and quite specific.  That is 
probably why the new generation working on 
orchestration/choreography are ducking away 
from real time requirements.

That is the conundrum of simplification and 
composability, also the inverse scaling effect 
of sign systems given operations. The web works 
precisely because it has few verbs and they 
are not coupled to the content.  But I've learned 
more about the problems of very large distributed 
systems from distributed simulations than any 
other endeavour.  It really exposes the specific 
problems when one is attempting to create, maintain 
and update a shared virtual reality in lowest 
common denominator platforms.   One combines 
lots of document/message types with just a few verbs 
for moving them and the trickery of delayed 
updates (why do the Matrix fighters freeze 
in mid jump: synchronization of views).

Yes, in general given both social issues 
and the energy costs of sharing interfaces for 
microdocuments, it is better to move coarse 
documents with rough agreements, ontological 
commitment, etc. in a lattice of potential 
meanings.  It falls apart in real time simulations 
though without a lot of trickery.  Real time 
latency is the bear if all viewpoints must 
share provably identical values.  The web is 
meant to be de-synced, so it just works.

Say 'batch'.

len

-----Original Message-----
From: Rich Salz [mailto:rsalz@datapower.com]

> I wonder who will be the first to resurrect Brad Cox and 
> Software ICs.  Now that is a golden oldie, but closer 
> to the solution than most.  It leads to C#, Java and UML 
> and away from declarative data objects.

Don Box is currently doing this as part of a talk about service-oriented 
architecture.  Services and documents are more flexible and powerful 
than interfaces.
	




 

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

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