I came across some interesting discussions on data models and data formats today on Vector One that are well worth a read: “What is a Spatial Data Model and Why are they Important to Understand?” and “What If A Global Modeling Environment Were to Offer Compatibility With Multiple File Formats?”
At Safe, we’re very familiar with the issues at play here. Over the years we have seen a number of initiatives come and go to derive an “uber-format” that can store and accomplish everything. The problem is that even if you can make such a thing, it becomes so complex that few tools will support it and even fewer users will understand it.
Jeff Thurston poses the question “Now consider – would we be any further ahead today if we had one data format?” In our experience, the answer is that it isn’t possible to even imagine a world with just one data format, precisely because, as Jeff acknowledges, “We need the right tools collecting the right data for the right purpose.”
Because of the diversity in purpose for geospatial data, there is a diversity in tools which have a diversity of requirements. This ends up imposing a diversity of file formats, each tuned for a particular situation and, even more importantly, holding a diversity of data models. It is precisely this data model issue, as the Vector One folks acknowledge, that truly is where the rubber meets the road when it comes to creating interoperating systems.
At Safe, we’ve been proud to play a key role in providing tools to help users smash through both the data model barrier as well as the format barrier. In our opinion, any attempt to get interoperability has to address both of these. Just the format is not enough.
As James Carville would say, “It’s the data model, stupid” …