Talk:New Proposed Architecture
From Intamap
--MattWill
iams 12:45, 26 March 2007 (CEST)
By providing the ability to pass observations in the actual interpolation request makes this solution more elegant than our previous idea but are there any negative aspects of this new architecture?
Excellent work
--Edzer 13:58, 26 March 2007 (CEST)
I'm very glad that this comes now, and not at the end of the project, for both SOS and WPS. I'm only slightly worried about the opening warning "This document is not an OGC Standard. ..." Though I'm convinced it will become one, do you have any idea as when we might expect that?
--MattWilliams 17:10, 26 March 2007 (CEST) I think we may have discovered them at just the right time. We have experience of using WFS/WMS to try and emulate what we want and now we can compare them to using a WPS. The main concern, however, is like you say it isn't (yet) a full OGC document. I'm not aware of how long it will take but will try and find out. I don't see this as too much of a problem though as there is a definate need for a "Web Processing Service" so can't imagine it being rejected.
--BrianDoherty 15:23, 30 March 2007 (CEST) Just for information, I introduced this architecture to Gerhard deVries and Konstantins Bogucarskis of EURDEP and contrary to what I suggested in Aston, they were quite positive about implementing an SOS for EURDEP. They could envisage a situation where a member state might access the Eurdep data directly through the SOS and incorporate it into their national system. It's therefore possible that this might be part of the actual EURDEP client implementation.
--MattWilliams 16:10, 30 March 2007 (CEST) Excellent. I will write up some more information about the Sensor Observation Service then. For development purposes 52 North is an SDK for Sensor Web stuff and can be found here. It is an open source project.
--Dan Cornford 22:20, 1 April 2007 (CEST) It looks fine to me. I have no major concerns about the architecture, indeed that fact that a WPS can accept any XML schema makes it more flexible in my eyes. I think the next step is to implement one as a prototype - Matt any idea of how long this might take? I hope it could re-use much code from the WFS?
--MattWilliams 10:29, 2 April 2007 (CEST) I think a prototype shouldn't take to long to create. Most of the code can be reused as the only thing that has changed (at this time) is the XML messages the service accepts - all the back-end interpolation code doesn't have to change.
