[
Lists Home |
Date Index |
Thread Index
]
From: Chiusano Joseph [mailto:chiusano_joseph@bah.com]
"Bullard, Claude L (Len)" wrote:
>
> The Justice XML schemas and data dictionary will
> have these in some fashion. How much actual use
> they are getting is questionable, but just as they
> did with CALS and 28001, Federal grant money is
> being attached to them for public safety systems.
>
> See www.it.ojp.gov
>The following URL provides information on "Organizations Utilizing the
>Global Justice XML Data Dictionary (GJXDD)":
>http://www.it.ojp.gov/topic.jsp?topic_id=107
Thanks! People should be well-informed.
>The first "official" release of GJXDD (it's also referred to as GJXDM,
>"M" for "Model") was a few weeks ago:
>http://xml.coverpages.org/ni2004-01-22-a.html
But is already referenced in an RFP sitting on my desk
as we speak and it is older than a few weeks. The antecedents
have been around for three years at least. I've counted
about four RFPs in the last eight months the reference
these. Of these, only one had them as firm requirements
and that one, some suspect, is written like that to
prove no vendors have it and therefore it is a good
idea to let the local IT shop do the work thus guarding
the jobs in cash-strapped economies.
>> Where the 80/20 point here is questionable. These
>> are designed for databases that have a high degree
>> of associativity among the elements.
>Not sure what you mean - could you please explain this?
I mean 80/20 is relative to the application. It's a no
brainer to extract elements for some application as long
as the scope of that is understood. It will be quite
another to critique the source without understanding
what it is designed for. IOW, one needs to understand
the data pipes of public safety. These aren't one
dimensional. One should read a lot of RFPs, know
what the agency scales are (eg, San Francisco is
not Pomona), know how diverse these are from agency
to state and from state to federal, know that these
same data elements don't really work globally, and
so on. Otherwise, it's a good model.
To understand
>> these, one has to understand the master index types
>> and how NIBRS is fed from local to Federal agencies.
>> One must also understand the differences among
>> operational and organizational statistics.
>Also not sure what you mean here.
This gets into a lot of detail. First, understand
UCR/NIBRS. Then understand these are not univerally
applied and are always locally customized. The high
costs are in the customizations and these are actually
necessary. So the Justice XML models end up being bid
as up/down translation targets and if the system is
wall-to-wall instead of best-of-breed, will slow the
system down and make it more expensive. Best-of-breed
will reduce the system's functionality so best-of-breed
ends up being 75% of the possible functionality.
Scale always means, pick a dimension. No free lunch.
len
|
