MAPSERVER: Implementation Question

[Umberto Nicoletti]

Neil,
if you already have all of your data in an Oracle database and you
have the license for oracle spatial then the simplest solution is to
use the Oracle spatial extensions.
In this way you can create spatial objects (points, polygons, lines,
etc) out of your data and then query them with mapserver directly.
Mapserver could first apply a filter (say what's that cost>3000 and
cost<6000) and then run the spatial query (say what's inside a
shape/rect drawn with the user interface).

HTH,
Umberto

On Dec 12, 2007 4:49 AM, Neil Atherton <Neil.Atherton at lmal.com.au> wrote:
>
>
>
>
> Hello All,
>
>
>
> I am currently part of a team developing a Task/Production Management/Work
> Flow Management application for a client here in Australia.  The core of
> this application is the concept of a "task", which is basically a set of
> requirements that come in from customers (our client's customers) and are
> passed around the organisation.  The result of a task is one or more
> "products", which are what is built in response to the customer
> requirements.  Both tasks and products have a geospatial component, in that
> they have coordinates stored against them.
>
>
>
> We have been asked to use Oracle as the DBMS and to build our application on
> top of TRIM.
>
>
>
> I have requirements to build a map search interface.  They want a map that
> they can draw a bounding box on and have the application return all tasks
> and products that fall within that area.  That geospatial search can also be
> combined with a textual search (e.g. date ranges, product author etc).
>
>
>
> That should be enough background…
>
>
>
> One of the solutions I've been looking at is Mapserver.  The web site says
> that data can be added via a delimited text file, and I think I've found
> examples of applications that use an index file to query against.  I had
> envisioned something like the following implementation:
>
>
>
>  ·        An index file (or files) is built when products/tasks are stored
>
>  o   This index file lists all products/tasks against a set of coordinates –
> it would be as simple as product/task number vs. coordinates.
>
>  ·        The user draws a bounding box on the map search interface deployed
> via Mapserver
>
>  ·        A query is made against the index file(s)
>
>  ·        The returned product/task numbers are passed to TRIM
>
>  ·        TRIM joins this search result with any other search criteria (e.g.
> date, author, title etc)
>
>  ·        TRIM executes this joined query
>
>  ·        The results can be returned to Mapserver somehow, so the user can
> see them graphically (e.g. a flag at each of the product/task locations)
>
>
>
> Does that sound sensible or is it too simplistic?  Our application really is
> light-on when it comes to GIS content, and the other solutions I'm looking
> at (e.g. ArcIMS/ArcSDE, Dekho/ArcGIS Server) look to be overkill.
>
>
>
> Thank you for your time,
>
> Neil
>
>
>
>
>
> Neil Atherton
>  Principal Engineer
>  Lockheed Martin Australia Information Systems and Global Services (IS&GS),
>  a division of Lockheed Martin Australia Pty Limited
>  Ph: 0419 035 319
>
>
>
>
>
>
>  ________________________________
>  This email and any files transmitted with it are intended solely for the
> use of the individual or entity to whom they are addressed. They may contain
> sensitive and/or inconfidence information that is LMAL proprietary
> information
>