SRC this morning announced it would make its Explorer geocoder technology available under an open source license. Directions Media editors offer their interpretation of the move.
From Joe Francica:
My take here is a huge question mark on what the reaction will be by the Business Intelligence (BI) solution community. SRC's approach in hoping to "democratize" location intelligence is a huge stretch. There is a basic assumption that by putting the Explorer geocoding engine into open source that it will spur adoption of other SRC geospatial analytical products like Solocast and Portfolio. That may be true, but I think BI vendors are having enough trouble understanding the basic business model for location technology integration as it is. While it seems that the BI vendors "love maps" they don't exactly know what software applications they should offer that truly utilize spatial analysis within the context of their own platforms. Their customers probably have enough trouble figuring out why OLAP cube analysis better supports ad hoc querying let alone worrying about spatial interaction models and psychographic profiling.
Dean Stoecker, president and CEO of SRC said in an interview that, "traditional vendors have made it hard to spatially-enable the database." This may also be true. However, while we need to expand the adoption of location technology with enterprise computing applications (e.g. customer relationship, supply chain, or mobile resource management), will an open source geocoder be that catalyst? SRC may benefit from its close relationship with Autodesk. SRC has used MapGuide as its preferred Web mapping engine. The coat tails of Autodesk and its recent decision to offer MapGuide as an open source solution may extend far enough to help drag SRC into the limelight.
I think SRC is assuming that the larger open source development community, not just the geospatial open source developers, will carry the message that the SRC Explorer geocoder is a useful tool. As a consequence, SRC hopes that CIOs and database administrators will look to other SRC products to see which analytical tools are worth considering for purchase. That seems to be the only way SRC profits from this action. Oracle may adopt Explorer as a standard geocoder; IBM may adopt it as well. Will a huge number of developers download the geocoder? Probably not, as this is just an engine, not a complete mapping application. This is not the same as the consumer-facing visual splash that Google made with Google Maps and its subsequent release of an API. That API, by the way, is not open source. And that's the rub. While the impact may not be immediately felt, any subsequent excitement in this action may be muted because the geocoder becomes simply an embedded technology, and who developed it may be moot.
From Hal Reid:
Geocoding, for most people is a done one address at a time; or maybe up to one hundred addresses by using a desktop mapping program. But for many enterprise users, geocoding means thousands of addresses, requiring an outside service or the purchase of expensive software to process those records. SRC has now made its enterprise geocoding technology, Explorer, free.
Access to a free geocoder, from the SRC perspective, could mean that a user will move out of the minutia of geocoding and into the real issues of Business Intelligence. Dean Stoecker disparages the well worn mantra of "80 to 90% of enterprise data contains some spatial element", and now believes he has provided the means to move beyond that mundane geocoding routine, to provide anyone the means to add coordinates to data so they can more quickly deploy the real applications of BI technology.
Of course, there is the possibility that this release may be a disruptive force, similar to that caused by original Wessex data in the 1990s. Wessex's lowcost street centerlines and boundary files caused significant expansion of Business Geographics. Possibly, SRC may "Extend-The-Reach" of Business Intelligence.
Consumers generally view geocoding as a commodity. It is free on Google, Yahoo, and MapQuest and is part of even low cost desktop mapping products. Without it, you can't locate your customers, the competition or even your house on a map.
If SRC is right, and mass geocoding is now a commodity, Explorer may be the enterprise equivalent. It may soon be ubiquitously embedded in industrial strength databases and BI tools.
From Adena Schutzberg:
SRC's decision to open source its geocoding engine is certainly good news for the developer community. While many free and not free Web services exist to do geocoding, I'm only aware of one standalone, open source US-focused codebase: Geo-Coder-US which powers Geocoder.us. That said, many an eyebrow was raised at the open source session at our recent Location Intelligence when an open source geocoder being developing in Japan was mentioned.
From what I understand, Explorer, as the engine is called, is not currently available from SRC as a product, but is an engine that enables its higher end analytical solutions. The company argues that making the engine available will enhance growth in the business intelligence sector since geocoding is the first step toward using the hidden aspect of location in business. That, SRC suggests, means more sales of its offerings.
That logic does not sit well with me. I have to compare it to the carefully argued case Autodesk's Gary Lang made supporting the decision to open source its MapGuide technology. One of the main arguments is that Web mapping is a commodity, or in common parlance, "Web mapping solutions are a dime a dozen." Autodesk (and therefore ESRI, Intergraph, et. al.) make little or no money on the software. That prompted Autodesk to offer the software free under open source license so it can set up a business model to garner profit via consulting, packaging and support of the free software.
But geocoding tools are not a commodity, especially standalone enterprise-level ones needed to support business operations. Yahoo and others offer free geocoding services that might be appropriate for "toys and mashups" but are not for "behind the firewall mission critical applications." Dean Stoecker made it clear in describing the initiative that only one of the key players in GIS (MapInfo, I believe) and none of the new mapping players (Microsoft, Yahoo, Google, etc.) and none of the database companies (Oracle et. al.) have their own geocoder. Save MapInfo, they license one or simply do not offer one. That fact suggests to me that geocoding is far from being a commodity, and SRC could make money on its technology! Said Stoecker, "geocoding is not a commodity, but it should be." I supposed now that SRC offers an open source solution one could argue all of the sudden geocoders are commodities.
While SRC makes its money in the business intelligence space (into which it grew after beginning in the micromarketing space), geocoding is needed across the board in GIS applications - including routing, environmental applications, emergency response, and in most government uses. Developers in all of those arenas will take advantage of this open source tool and while they may contribute code back to the project (the software is licensed under Lesser General Public License, LGPL) it's unlikely that use will encourage sales of SRC products.
All that said, I respect SRC's decision to make what many will likely call a bold move. One of the privileges of running a privately owned company is that such things are possible with little or no discussion. But, sometimes that means proper scrutiny is not ensured. From what I understand, Autodesk held up some pretty high hoops for its open source proponents to jump through before approving the MapGuide deal.
Besides the altruistic and weak "it'll drive more business to SRC" arguments, then, why did the company decide on this path? Market disruption? Disruption for disruption sake is not a business plan. Is the goal to tap out a revenue stream for the single competitor in the "GIS space" with a geocoder? (QAS, Group 1 and others offer geocoders, but Stoecker puts them in a different category.)
Is this a marketing play? Is being the second commercial geospatial company to turn something open source worthy of press and recognition? Sure, and we are proving that by writing about it here! Still, GIS is a host of niches and while the geocoder appeals to many disciplines that use GIS, SRC products will only appeal to a small number, so the marketing value, at least in my mind, may be quite limited.
So, on balance, I'd give "thumbs up" from the open source community which now can point to a well tested open source geocoder. If I were a shareholder of SRC (it's privately held, so I cannot), I'd give thumbs down, as I'm not convinced the share price or revenues would rise.