[Webmap-discuss] GeoMoose vs. OpenLayers for specific project (Was Re: [mapserver-users] Advice for MapServer project)

[Christopher Schmidt]

This discussion carried over from MapServer-Users list. This is about
the benefits GeoMoose offers over OpenLayers.

On Tue, Feb 26, 2008 at 05:21:13PM -0500, Christopher Schmidt wrote:
> On Tue, Feb 26, 2008 at 04:12:58PM -0600, Bob Basques wrote:
> > The setup of the interface as well as the data sources are all 
> > abstracted to the point that each individual data custodian, and/or 
> > service owner, can implement their respective dataset/service in their 
> > own way and update it as frequently, or infrequently as they desire.

To me, this sounds like server-side layer management: is that what
you're talking about?

> > Everything in GeoMoose, including the Nav tools, are considered a tool 
> > or service of some sort, so it's very easy to add others to the mix.

These sound like OpenLayers Controls: Attribution, OverviewMap, etc. Is
that what you mean? Is there something special in how GeoMoose does
this?

> > The administration of the data maintenance is also a big difference 
> > compared to other pacakages, because once the layer/service  is 
> > implemented via GeoMoose, it's up the layer/service providor to keep it 
> > maintained into the future.   Since everything is a service of some 
> > sort, these services can all be maintained and authenticated in any 
> > manner the custodian/owner want to use.

Is this server side code? Client side code? This sounds sort of like
OpenLayers 'layers'; I'm not sure how it's different, other than that we
maintain the layers in the OpenLayers code. Does this make sense? Do you
do this differently?

Is WMS a 'layer/service' in GeoMoose?

> > GeoMoose is aimed at datasets that change fairly often, on the order of 
> > hours or minutes, and is intended primarily to access this type of data 
> > set.  

Sure: how does the client accomodate for that? Automatic refreshes?
Cache busting? Is this configured through the XML?

> > OpenLayers with it's Tiled interface seems to be intended for 
> > displaying datasets that tend to be more static in nature.  

This may be true for some people. Of course, OpenLayers supports:
 * Automatic merging of 'cache busting' parameters into the layer
   parameters via layer.reload
 * Untiled interface to layers as well (just turn on the singleTile
   option for layers which support it) 

And of course, in my mind, the time length for which data is valid is
not related to the client, but to the *server* which should deliver
appropriate caching headers: OpenStreetMap does this for its
'Osmarender' layer, which will automatically refresh tiles which are
updated as you drag the map around.      

> > Now, either of the two packages can can be used for each others
> > capabilities, but GeoMoose was built from the ground up based on a
> > asset management type of business need with mapping as the visual
> > identifier.

Hm, I'm not sure I understand this: perhaps this is a key point. Can you
describe "asset management type of business need" more clearly? 

> > One last item, is that GeoMoose the interface is initiaqlized from a XML 
> > file, a MAPBOOK, so setting up business specific interfaces is very 
> > easy.  

I'll admit that OpenLayers eschews non-standard XML-based configuration,
tending instead towards standards like WMC for configuration via XML.
Mostly, howeveer, it's true that OpenLayers expects users to configure
their application via Javascript. MapBuilder is the tool to use if
XML-based configuration is really your thing, and you're interested in
OpenLayers: It's all done via some XML document that is being
standardized as they build it or something, as far as I understand it.    

> > This XML file is even something that can be automated, so that an 
> > online method of setting up a Mapping interface could be built, and is 
> > something that is planned for as well.

Sure, but you can build that without XML :) Lorenzo Becchi has been
doing that with his WMS capabilities reading layer loader thingy, I
think. It then serializes to JSON or XML context documents, 
which can then be reloaded later on. 

Looking forward to learning more,
-- 
Christopher Schmidt
MetaCarta