Re: [projectvrm] UK gov launches "midata" data giveback program

[Brian Behlendorf < >]

I suspect that the first stage of midata will result in data dumps from companies without much metadata.  Your Tesco data dump, for instance, is likely to look very much like a compilation of all your receipts from the Tesco cash register.  A glance at a typical receipt these days will show a ton of fairly inscrutable terms and numbers, from which you might clearly make out the total and date of transaction, but might struggle to figure out what it was that you actually bought.  I suspect a secondary set of services will evolve... will need to evolve... that help parse all that into something more semantically meaningful.  That parse algorithm should be common property, a la wikipedia, so that they can be more quickly built and continually tuned, and so post-parse data could be shared between PDS's.  Perhaps this is a useful project for the Customer Commons.

Brian