Visualizing the Information Space

Things are quiet and this is Friday, so this is cross-posted 
from some notes I've made to XML-Dev and elsewhere.

If you ever find yourself in front of an audience trying 
to explain the why of XML, particularly to programmers, 
try this:

XML interoperability is often misunderstood because since XML 
does not have operations, what is it doing to make a system 
interoperable?  The network effect? The network effect is a 
power law, not a force for good or evil. Like XML, it does not care.

Think about it as two dimensions of interoperable scalability.   
In the x, there is data.   In the y, there are operations.
The x dimension scales almost infinitely if the names are well chosen.  
The y dimension does not scale well as most operations are local.  
However, when the x names are chosen well, that is, are 
semantically potent, the x and y dimensions couple via the network 
power law to drive the value of x and y simultaneously.
 
That coupling is why XML won.

If you plot that, the number of possible users of 
a data or operation is the Z axis, roughly, the range 
of the network power law.   If you plot against 
real data and users in real time, you should get a nice 
cluster animation.  The local clusters should roughly 
correspond to application languages and users.  Might 
make an interesting desktop for application grids.  Near 
functions by roles and language overlaps represent families 
of applications and application users.

That is a lot less boring when animated then explaining 
an information space and why the web architecture works.

len

Received on Friday, 13 February 2004 15:40:06 UTC