Tuesday, May 14, 2013

Rift in GIS ranks

There is a troubling development in GIS community that threatens to flare up old animosities and split participants along the lines of big versus small as well as proprietary versus open source. In the centre of it are Open Geospatial Consortium (OGC) and its members, representing open source camp, minor proprietary solutions vendors and 800 pound gorilla in GIS – ESRI. Ultimately, the whole incident can undermine, one way or another, OGC stronghold on all things spatial…

Firstly, a short history of OGC, from my perspective. In the beginning OGC was just an obscure “public benefit” organisation focused on developing and promoting open and interoperable standards to facilitate exchange of spatial information. There was a real problem that required urgent attention – that is, finding an easy way to exchange vast amounts of valuable spatial data stored in “walled gardens” and in obscure proprietary formats.  So, many were prepared to put a lot of effort in their spare as well as “employer paid time” in order to come up with a solution to this problem. This early period attracted a lot of prominent participants from government organisations and academia across the world, as well as support (often financial) from several industry sectors hoping for significant cost savings in the future. It was a community wide initiative with roots firmly established in an open source movement.

When OGC standards started gaining traction and wider acceptance amongst key buyers of GIS technologies the organisation attracted to the cause members from a proprietary solutions camp as well. OGC reached then the peak of its relevance.  A new era started where a united GIS community worked together towards a common goal – achieving full interoperability of spatial data, regardless whether data originated in proprietary or open source formats, or using proprietary or open source tools.

However, the relevance of OGC and its initiatives is under threat now - from outsiders as rapid technological advancement demands more flexible, easier to implement and performance focused standards (like those developed by organisations and individuals operating on fringes of GIS  - Open Street Map initiative and GeoJSON are the two most prominent examples), but also from within its own ranks as the organisation is forced to deal with conflicting interests of its members.

In particular, ESRI had put forward a new GeoSpatial REST API standard for endorsement by OGC. It covers a wide set of web services including catalogue, map, feature, imagery, geometry, geoprocessing, and geocoding web services.  However, a significant number of OGC members lodged an objection to it and demand the documentation not to be adopted as an official OGC standard.

The essence of the objection is that the proposed standard is not a community designed scheme but rather just a documentation of an implementation of proprietary web services by a single commercial vendor.  No other vendors will be able to build solutions against that standard, and if endorsed, it will dilute relevance of existing OGC standards.  When it becomes an OGC endorsed standard there is a danger that buyers will demand that all the vendors of spatial solutions support that standard in their products, hence giving an unfair advantage in the market to vendors of ESRI solutions. The objections are fully documented on OSGeo website.

Unless a workable compromise can be found, then regardless whichever way OGC decision goes it will alienate a significant group of its members and ultimately, it will split GIS community. That is, it will upset either open source and non-ESRI vendors or ESRI and thousands of its affiliates whose livelihood is almost exclusively dependent on ESRI’s fortunes.  If things get rough ESRI is dominant enough in the market to go its own way, without the blessing from OGC. But will OGC be able to retain its relevance without such a prominent member as ESRI? Potentially some troubling times ahead for GIS industry...

[Update: ESRI has withdrawn the GeoSpatial REST API standard from OGC, potentially defusing the situation. ]

No comments: