I and the OS are great supporters of GML, the Geographic Markup Language, an open standard for XML encoding of geodata. OS MasterMap was the first widely available product to make use of GML and over the past couple of years the industry has developed solutions to fully exploit it.
We knew when we started using GML there were some limitations and our application schema had to extend the then core schema to account for the richness of MasterMap data – this is a great advantage of GML, the ability to extend the schema to meet the needs of a particular user community or supplier is a great advantage, if vendors correctly (and many don’t!) implement their GML parsers it is a very flexible solution.
This flexibility however has also been GML’s downfall to this point, many perceive GML to be very complex, verbose and unfriendly – and I must admit to having a little sympathy for this view, as would anybody who has looked through the GML 3 documentation all 800 pages of it !
But key to the solution here is the L in GML, the fact that GML is a Language, means when can select just enough of the language to communicate what we need to and ignore the rest.
For a French speaker, although English is a very complex language, you need only a subset of it to communicate while on holiday, likewise to communicate simple geodata you need only a subset of GML. The idea of simple subsets of GML or profiles has been bouncing around for at least a year, so it is very good news to see that the OGC has published a candidate specification and is asking for public responses.
I can’t see geodata web services taking off without profiles becoming adopted.
Written and submitted from home, using my home 802.11 network.