RE: [emergency] Re: Circle and Polygon

["Bullard, Claude L (Len)" <len.bullard@intergraph.com>]

It's simply two things:

1.  With the reference element, the software has a 
standard way to notice different systems.

2.  Without the reference element, the different 
systems can't interoperate.

That's not rocket science.  It's JIT conversions 
vs dark spots in the coverage.  You have to decide 
if the reduced performance of some nodes is worth 
including them in a network of responders for which 
the majority have optimum response times and the 
minority will evolve predictably toward the optima. 
That is, better some than none.
 
What you want to know is the cost of that given 
different approaches and the actual impedance 
mismatches (semantics and format) of the different 
systems.  XSLT is cheap but would you run that 
on the sensor?  No.  You would put that in the 
smart aggregators that sit between the sensor 
and the consoles.

The economics are a little less clean.  If a vendor 
proposes that the system not enable standard 
software to notice different systems, then 
goes to a local market and *discovers* that the 
local procurement requires support for the 
legacy system, that vendor will add a proprietary 
extension to the standard to enable that, thus 
locking in the customer to the proprietary solution  
for the lifecycle of the system.  The lifecycle 
of public safety purchases is approximately 12 years.

len


From: Rex Brooks [mailto:rexb@starbourne.com]

Offhand, I would say that including validation checks as a mandatory 
item in RFPs also makes sense to me. Assuming that good practices 
will be followed sounds like a train wreck waiting to happen. Making 
demonstrable, and accurate interoperability with a given set of CRS 
might give a few vendors heartburn, and make some project managers 
grumble, but, while it isn't trivial by any means, it aint QUITE 
rocket science or brain surgery. It may be just a bit more important 
at the time of response coordination, though.