Christopher’s Priorities

Christopher Gutteridge’s post is at

The revolution will be interoperable. Christopher Gutteridge describes his vision for the utopian future of digital information

1 comment

  1. Stephan Kreutzer - January 12, 2018 6:43 pm

    There’s a way to make our and foreign data meaningful by adding semantics, and it’s called XML. It’s enough if the MIME type states “text/xml” (of course we have to see if a general XML file handler will respond to it and pass the data to a specific client/agent/application upon payload/content inspection, or if we just add new MIME types), and in the XML itself, the DOCTYPE declaration or namespace states what semantics apply. As it is up to everybody to come up with their own XML-based format and as it’s up to client implementors to decide if they want to interpret the markup according to the official definition, eventually there can be a lot of competing, coexisting and cooperating “standards” (small and big ones), but as it doesn’t make a lot of sense for a format about product prices (fish) to specify location semantics (vendor), they need to be intermixable, which already works today based on the notion of microformats and using namespaces on sub-elements. True, it can get a little bit messy when dealing with such intermixed data (if it is one big hierarchical file containing everything, for example other worksheets of the spreadsheet application we’re not interested in), so we might want to translate those XML conventions to more Doug/Ted/NOSQL-style data structures, where all data stays isolated for itself, and we interlink it via generic conventions that don’t lack a way to bind semantics to the referenced data. This should be easy and fine for mostly static documents, for raw/application data, I’m not so sure as developers may change the semantics frequently without updating some kind of WSDL/XML-Schema for it, which led to approaches like ReST at least in theory. There are plenty of people doing some part of this already, but it’s usually none of the big Internet companies, because if a generic solution would be rolled out at one point, they would have to change their business models as everybody would be able to organize the world’s information as a curator and knowledge worker.


Have your say