Re: [egov] XML Schema Design and Management Guide for Hong Kong SAR Government

[David RR Webber - XML ebusiness <Gnosis_@compuserve.com>]

Team,

I'm forwarding here the dialogue from over on the
OASIS eGov list.  As you can see this started out
as an interesting thought - and has developed into
interest in looking at CAM to make GML more 
accessible / quicker to do.

If anyone is specifically interested in working
on this please let me know.  I'm planning to 
liaise with Carl and provide him with PPT and
assist in understanding what CAM can do today.

This is a complex area and theoretically this
will no doubt test our model significantly!

At the very least we may find out possible
future V2.0 thoughts - but obviously if we
can deliver useful value with V1.0 right now
that will be great.  At this point I'm not
even sure what V2.0 is - but its great to have
different sets of folks validating our current
work!

Thanks, DW

-------------Forwarded Message-----------------

From:	"Carl Reed", INTERNET:creed@opengis.org
To:	"Eliot Christian", INTERNET:echristian@usgs.gov
	"David RR Webber - XML ebusiness", Gnosis_
	
CC:	"Kurt Buehler", INTERNET:kurt@opengis.org
	
Date:	7/25/2003 11:23 AM

RE:	Re: [egov] XML Schema Design and Management Guide for Hong Kong SAR  Government

 
David -

Thanks for the detailed response! And I like your suggested way forward
"Sets of CAM templates that illustrate for common GML constructs exactly
what the semantic markup composition is". Perhaps this is a good area for
collaboration. If so, then we need to bring some of the OGC member GML
guru's into the discussion - would make your life a bit easier :-)

Regards

Carl

----- Original Message -----
From: David RR Webber - XML ebusiness <Gnosis_@compuserve.com>
To: Carl Reed <creed@opengis.org>; Eliot Christian <echristian@usgs.gov>
Cc: Kurt Buehler <kurt@opengis.org>
Sent: Wednesday, July 23, 2003 8:08 PM
Subject: Re: [egov] XML Schema Design and Management Guide for Hong Kong SAR
Government


> Carl,
>
> You no doubt saw my other posting!  Your work took my breathe
> away - I had no idea that someone had built such a world view
> for themselves.
>
> I was project manager for the USDA geospatial team for a while,
> and I know Eliot Christian here in the USA from the GILS team.
>
> My current obsession is with VisualScripting - http://www.visualscript.com
> which is the opposite polarity.  How to make stuff simple to do, by making
> the tool do the work.
>
> When I saw your post on GML I thought - wow great stuff for
> VisualScripting.
> I came away licking my wounds!  Well maybe next year when we've added
> a chunk more to the toolset.  Right now the best we can do in this
> direction
> is ontologies for topic maps, and of course the floorplan example style
> stuff.
>
> Anyway - enough rambling - back to your question.  Hmmmmm. The
> OASIS CIQ work is generalized addressing - so forinstance the Corp of
> Engineers could use it to address weirs, dams, locks, pumping stations,
> and that ilk.
>
> Now using VisualScript I've developed this cunning technique of when
> something is dropped onto a map overlay (so it could be a symbol of
> a pumping station, and the overlay could be a satellite image of the
> terrain) - then it can inherit the GPS coordinates from that layer without
> the users having to key that in.   So now - you could associate the
> terrestial address - something like Pump #21, Dam Road 157, District 17,
> South Mountain Runoff, Garrrett County, West Virginia - and that format
> could be driven off the fact the icon is a Corp of Engineers pump, and
> that's how they address their stuff.    And you could use CAM to define
> those context rules.
>
> Whenever I get this deep into the markup however I like to step
> back and ask - just what business problem are we trying to solve,
> and does anyone care enough to worry about this?   Afterall - if I have
> the GPS - I can throw someone in the front of a 4WD truck and let them
> bounce around for an hour till they find it, and a route to get to it!
>
> So - I have to ask - assuming we could tie context into all this -
> what good is it?!?
>
> Maybe its just the markup itself?  Looking at GML - I'd have a really
> hard time figuring out for a given set of information - just how to
> present and store that in the XML so that someone else can make
> use of it.   So - is this the holy grail?  Sets of CAM templates that
> illustrate for common GML constructs exactly what the semantic
> markup composition is - and how those nouns and their use tie
> into the overall world view?
>
> If that's the answer - then - yes - we can do that bit handily!
>
> Cheers, DW.
> =======================================================
> Message text written by "Carl Reed"
> >David -
>
> Interesting email. Very informative. I really appreciate your cautionary
> note, "that if you create a W3C Schema XSD definition that you have
created
> information interoperability." In the OGC, we have a specification called
> the Geography Markup Language (based on XML and XML Schema). Our members
> defined GML knowing that it solves only part of the interoperability
> problem
> of spatial data encoding and communication (sharing). The other part is
how
> to express physical content models and semantics in a well known and
> useable
> manner. As such, we have been recently working the area of schema mapping
> and domain modeling.
>
> Now, I do not know if this related to your work at all as I am not
familiar
> with the work of the CAM TC. However, it is interesting that there appears
> to be overlap between the work of the OGC membership and the work of the
> CAM
> TC. We have done considerable work on specifications (along with ISO) for
> metadata, registries, catalogs, information sharing, and so forth. All of
> this work has a spatial focus but we draw heavily on the work of other
> standards organizations. I was especially interested in your example of
> addresses. OGC members have done much work on interfaces for what we call
> geocoding (a service that transforms a text string, such as an address,
> into
> a x,y coordinate with a set of associated properties). As a result of this
> work, we have done a considerable amount of work in the area of "address"
> as
> an abstract data type. However, we did not deal with context driven
> assembly
> and so forth. I therefore believe the work you are doing could be very
> applicable to the work we are doing.
>
> I may be way off base, but I thought I would ask anyway :-)
>
> Regards
>
> Carl Reed, PhD<
>