Open source crusade blocks geospatial standard

An open standard proposal by mapping giant Esri has failed after a backlash from developers within the geospatial community led it to withdraw from the process.

Esri, with around 30 percent market share in geographic information systems (GIS), has over one million users and 350,000 customers, most of whom use its ArcGIS software.

ArcGIS has -  for many users - become a de facto industry standard for creating maps, matching those maps with geospatial data, and using the results in software applications, in much the same way as Microsoft’s Office dominates word processing software.

In early 2011, Esri submitted an API (application programming interface) for ArcGIS (known as the ‘GeoServices REST API’) to standards body the Open Geospatial Consortium for consideration as an ‘open standard’.

Esri proposed that because its GIS server adhered to existing web services standards and because use of the API was freely available via the Open Web Consortium (OWC), it could - much as Microsoft managed to for its OOXML office document format - have it considered an ‘open standard’, all but guaranteeing Esri’s market influence for the foreseeable future.

For a time, it appeared Esri had the full (and somewhat unorthodox) preliminary support of key influencers within OGC standards body. An Esri-branded presentation encouraging the geospatial community to back the proposal, presented in February 2012 and sighted by iTnews, listed OGC president and CEO Mark Reichardt on its opening slide. Reichardt has since informed iTnews that he was listed as a panelist to discuss the proposed standard and was not a co-author, as previously reported.

This presentation described the specification as “a proven and easy to understand method for a broad range of clients and applications to request map, feature, attribute, and image information from a GIS server.”

The OGC - which takes in the views of some 480 member organisations - had long before approved open standards that would achieve much the same result (such as WMS, WFS, WCS, and WPS). But Esri argued that its API reflected the changing needs of today's software developers. Being a JSON-based, RESTful specification, Esri claimed the specification would “make the GIS server instantly usable by thousands of developers working in popular client-side development environments.”

What Esri, its users and partners failed to predict was the backlash from a large and vocal group within the geospatial community that viewed the proposed standard as being in the best interests of a single commercial entity.

Protest

Australian software developer Cameron Shorter, the local chair of the Open Source GeoSpatial Foundation, a group which specifically lobbies for the use of open source technologies within the geospatial industry, was among the loudest voices raising concerns over Esri’s candidate (proposed) standard.

Shorter has spent the last few months attempting to disrupt or influence the process - either by writing blog posts or by contacting key OGC members that may not have realised the ramifications of adopting the standard.

Shorter argued that while Esri’s technology and interest in open standards was laudable, approving such a specification as an open standard would do little to improve on existing, genuinely open standards that achieve the same result.

Chris Holmes, a fellow open source advocate, predicted in a blog post that many in the industry did not trust Esri's motives. Holmes worried “[Esri] would embrace the language of open, but not actually do the key things that would make a real open ecosystem” , and that most of what the software company offered as 'open' was “useless without pieces of closed software.”

Shorter wrote to government officials with voting power on the OGC warning them of what he perceived as Esri’s loose definition of ‘open’. Supporting such a proposal would not result in interoperability - the ultimate goal of standards, he said, but the opposite.

Shorter explained to iTnews that dominant vendors often wish to have their proprietary interfaces nominated as ‘open standards’ as government buyers in particular have wised up to the adverse effects of vendor lock-in. 

Should that result in “incompatible duplication” with existing standards, it creates a higher cost for other parties’ solutions to be compatible with both standards, and thus puts existing ones at risk.

In the wash-up, he said, mandating a standard put forward by a commercial vendor is tantamount to a government “providing a vendor with significant market advantage, erring on the creation of a state-sanctioned monopoly."

Read on to find to how open source advocates rocked the vote...

Rock the vote

After two years of discussions, the OGC invited members with voting rights to vote on the candidate standard in early May 2013, but the process stalled within a matter of days in the face of several ‘no’ votes.

Behind the scenes, both advocates and opponents of Esri’s proposed standard oiled up their lobbying machines, despite OGC overtures to avoid any inflammatory activity that would threaten the progress of the vote.

Esri employees, wising up to the opposition swelling around their proposal, took the extraordinary step of personally lobbying customers that were OGC voting members, insisting that any objections were without merit.

Shorter, meanwhile, grew increasingly concerned that government users of Esri software would not have the time to investigate the matter fully, and that many would yield to pressure from internal users wedded to the ArcGIS product. His fear was that while a ‘yes’ vote from these organisations would see Esri’s proposal approved, any decision to abstain from voting would also invariably tip the numbers in Esri’s favour.

Shorter embarked on some last minute lobbying of his own. A broad mix of competitors, open source advocates and concerned government officials  - most of whom were OSGeo members - were asked to consider signing an open letter to the standards body expressing their objections. This was eventually signed by some 135 individuals and stated a list of reasons why Esri’s proposal should be rejected.

These efforts now appear to have borne fruit.

Last week, a member of the OGC’s Standards Working Group (SWG) passed a motion to withdraw the specification from the OGC standards process - a motion which was unanimously accepted. 

If there are no objections from the OGC’s Technical Committee before Friday, June 7, the specification is withdrawn. (If there are objections, the motion becomes the subject of an electronic vote.) A representative of the OGC has told iTnews there have been no objections to date.

It has since been revealed that the SWG member that put the motion forward was an employee and official representative of Esri (Keith Ryden), and the motion was seconded by a key Esri business partner.

This has even Esri's most staunch critics wondering whether this was an act of grace or cunning.

Representatives from Esri have been contacted on several occasions by iTnews to ask why the company chose to withdraw the proposal, to no avail.

The wording of the SWG’s motion offers some clues.

“Considering the breadth of discussion both internal and external to the OGC process since the vote announcement, the SWG members feel that the vote cannot continue until the many questions raised have been addressed,” the SWG wrote in the motion to withdraw the Esri specification.

“Issues regarding OGC process, vendor advantage, duplication of capabilities, etc. have now overshadowed technical discussions of the merits of the specification. By withdrawing the OGC GeoServices REST API candidate standard, the necessary discussions regarding OGC process, policy, and position can continue separately.

“Over the coming months there will be opportunities for OGC staff, members and the general public to engage in discussions related to policy and procedures, such as clear statements on openness and interoperability, overlapping standards, backwards compatibility and so forth.”