With the arrival of OS Open Data I thought I'd dust off my old skool data skills in an attempt to create the first derived dataset from OS Open Data. Using our GO Publisher product I mashed OS Code-Point Open with the WOEIDs of Yahoo! GeoPlanet Data and created a derived product in KML , here's the detail ...
1) First things first, make sure that historic open data licence is covered in the header ...
2) Convert to KML (something a little more internet friendly than the default CSV)
3) Sort out the spacing on the Code-Point Open Post Code making it single spaced.
4) Convert the coordinate system to WGS84. Note: Where the Positional Quality value of 90 (geometry unknown - usually 0.0, 0.0) there is NO Point geometry. You'll notice that the Post Code value is in the kml:address tag and so if you display in Google Earth it will geocode any missing geometry on the fly for you.
5) Put all the Code-Point Open properties in the kml:ExtendedData tag
6) Process the Post Code to create a new Out Code property also in kml:ExtendedData
7) The cool bit ... Cross reference the post code to Yahoo! GeoPlanet Place postcode, attaching woeid and parentWoeid to the original OS Code-Point Open data.
All credit goes out to our GO Publisher Community Edition product which let me do all of this so quickly. Here's some screenshots and there's some sample data for comment at the end of the post.
You can access the first cut sample data here:
http://www.snowflakesoftware.com/webservices/opendata/cpoww-sample1.kmz
Just click on the link and it should open up in Google Earth, if not open Google Earth first then click on the link.
I'll put this sample out for comment, if people are happy with it I'll create a national set for you all.
Enjoy
Ian

I, for one, would find a national map of constituency boundaries (for UK, Scottish, Welsh and Northern Irish elections along with European and maybe even more localised elections) in a format that can be used with google maps, very useful.
Please keep up the good work.
Thanks for the positive feedback Garry.
We've actually just released a guide on how to publish to KML so you might find that useful too:
http://blogs.snowflakesoftware.com/techies/2010/04/publishing-kml-using-go-publisher-community-edition.htm
Hi Ian, look good! We did the conversion of all the post codes to xml format. We are using it to geolocation in one of our softwares. The xml file is used to locate boreholes positions on Google Earth (import/export)