harvesting

Norman Barker nbarker at ittvis.com
Thu Aug 3 06:29:46 EDT 2006


Hi,

I have a crawler that looks for getcapabilities (or service) references
from a given input url, it doesn't go any deeper than the first page at
the mo' (lets just get it working!).

Now, OGC says that the service model should be common amongst the
getcapabilities so that we have 

<ows:ServiceIdentification>
<ows:ServiceType>.... Etc

However as we know just from looking at OnEarth or ICEDS that they tend
to be service specific, so that wms has keywordlist, wcs has keywords
the list goes on.

Once the crawler finds a service it passes the input xml doc through a
stylesheet to process it into simple metadata kvp, so contactname=...,
service=..., keywords=...

Is it fair to write the application so that it matches on the ows
namespace, or shall we be backwardly compatible and parse on version,
and if so what minimum version number shall we support?  The stylesheet
approach is language neutral as well.

Norman




More information about the Geodata mailing list