Excellent - better to solve some hairy problems once than spend your resources fighting over alternative flawed implementations.
Apropos of this, its time to point the community at some thinking coming from the internals of an ecosystem: driven by business cases and data models, as well as experience and demand for interoperability:
see: http://www.metoffice.gov.uk/informatics/index.html
Note in particular the concepts around "declaritive semantics" - in this case "processing Affordances" - ie. abstract interfaces that data models can support in order to allow processing via particular processes.
This rather elegant unifying concept allows us to see simple features profiles of GML as simply views that allow 2D geometry operations, and coverages likewise as one possible representation of a "field" of values (in n-dimensions). This also makes WCS/WFS/other data services reconcilable at the semantic level, and thus makes it easy to deploy WPS to operate on (or transform between) different representational views. For example, a contouring service could take a 2D grid and deliver a series of contour features. Another could take a 4-D gridded model output and create a time-series of contours using the same data model for contours, even though different properties could be attached to the contours.
To take the collaboration forward, it would be useful to have a target "business architecture" that doesnt warp the solution to be too specific. This is always a major challenge: how to get the right perspectives around the table during the early design phases.
One idea might be to set up an "open architecture" mailing list for any such discussions. A range of projects have been using the Wiki at https://www.seegrid.csiro.au/twiki/bin/view/Main/WebHome to collaborate internally and make common elements more visible. This site would need revamp, maybe a ReferenceArchitecture wiki that pulls these common elements into focus. The "Infoservices" wiki has a preliminary roadmap that needs extending, even though its prognostications (two years old) are all perfectly valid, there is much to be added to make current project scopes more easily understood in terms of the high level view of the requirments.
Rob Atkinson
Chris Holmes wrote:
Hey everyone, in the past couple weeks The GeoServer Project has become aware of FROGS (framework for open geospatial services, http://frogs.tigris.org), and the WPSInt plug-in http://wpsint.tigris.org/. Their team has built a solid framework to support new versions of OGC specifications, and already have one success with the implementation of the Web Processing Specification. What they lack is a solid open source community around the framework. As GeoServer is looking in a similar direction, for what we are calling GeoServer 2.0 (http://docs.codehaus.org/display/GEOS/GeoServer+2.0+Technology), we've decided to combine forces.
The plan is that we can collaborate on a base framework that supports pluggable services for all our current needs and designed with future ideas in mind. We will be able to work together on the base level and have more time to create new services and improve existing ones instead of replicating work on all the more mundane tasks like supporting different data formats, J2EE integration, data configuration, and eventually user management and security,. Together we hope to become the framework of choice for web based geospatial services. This will likely eventually spin out as its own project, with 'GeoServer', 'WPSInt', and others being 'distributions' of plug-ins on top of the same base framework. We hope to support a diverse ecosystem of communities around web based geospatial projects, just as Eclipse does in general and uDig is starting to do with desktop GIS.
There are no immediate plans to perform this work, but our developers are starting to get acquainted with one another's code, and we will point newcomers at one another's respective projects. The hope is to find some funding to be able to dig in to the work of combining the projects relatively soon, or to at least build it in to another funding opportunity. But even if that doesn't happen, at the very least we will orient the directions toward one another, and if it can't happen immediately we can work there over time. We are all excited about the possibilities.
best regards,
Chris