[postgis-users] Re: FGDC/ISO ... Metadata
ymoisan at groupesm.com
Tue Jan 18 08:18:29 PST 2005
Hi Homme !
Thanx for your input. Using a blob to store the XML and the application to do the validation certainly will fit the bill, but I would prefer to rely on built-in validating parsers and XSLT engines that usually come with XML backends. I fear the changing needs of my users to store different metadata elements (both among the users themselves and as they learn about metadata!) would make a solution like the one you're proposing a maintenance nightmare in my case. I'll try to shoot for a maximum of (meta-)data related logic to be carried out in the backend and leave my front-end app to do things like user/document management.
Another point is that with the CGDI zeroing in on ISO 19115, there is an obvious general need for an easy way to glue metadata information to spatial data. A lot of people can develop their in-house solutions, but since it's such a general issue there is a definite need to have easy paths to doing it. I was thinking of an XML backend because you can supply a subset of the metadata elements and still use the backend built-in utilities to validate it and store it efficiently. It's a way to save space just as you did with BLOB's (as opposed to having this huge 95%-empty table/field structure that would need to be created to accomodate a model like FGDC-CSDGM), but it is more amenable to manipulations IMO.
The application we will be developing for example involves using water quality sensors. I want to be able to link the data with the sensor specs (e.g. SensorML) that will provide all sorts of interesting things like data validity domains which can be used by data integrators afterwards (e.g. to assess whether the data meets the requirements for input in a particular model). Also, there will be people uploading data they got from their personal pocket sensors, so again a bunch of metadata.
Thanx for reminding me abour XpXML. Doesn't look too strong developmentwise, but I'll give it a look.
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the postgis-users