GEOG 583
Geospatial System Analysis and Design

Reading Assignment


For this week, I'd like you to read two pieces on GIS and open source, and watch one video. The first article is by Arnulf Christl, on open source business models. Arnulf runs the WhereGroup GmbH which supports open source geospatial products. The second is an academic paper that looks at open source GIS tools by Stefan Steiniger and Geoffrey Hay; they are landscape ecologists, but the paper seems relevant for most types of geography, too. Finally, the video by Paul Ramsey of OpenGeo is on the value of open source.


Free Software and Open Source Business Models by Arnulf Christl, a chapter in Open Source Approaches in Spatial Data Handling by G.Brent Hall and Michael G.Leahy (eds.) (2008).

Think About

While you're reading, I'd like you to consider just how Arnulf's ideas could apply to your work; could you or your organization have saved money using open software or open data sources during the last year? What other (non-financial) costs and benefits should be taken into account when considering open source software and open data sources?


The following video is a keynote is from Paul Ramsey of Boundless, who is a co-founder of of the open source spatial database project PostGIS (21:28).

FOSS4G 2011: Paul Ramsey keynote from Peter Batty on Vimeo.

Click for Transcript of FOSS4G 2011: Paul Ramsey keynote

SPEAKER: Thanks a lot, [? Arlov. ?] Next up is Paul Ramsey, who I've mentioned in passing. So Paul is in many ways responsible for me being here today through being kind to me up in Victoria four years ago, I believe it was. And Paul always does a great presentation. We call this Mr. [INAUDIBLE], really. So Paul.

PAUL RAMSEY: Thanks a lot, [INAUDIBLE]. So why we do the things we do? Oh, in open source business models. Lately, I've been teaching my daughter Pig Latin. Actually, a family variant called Arb, [SPEAKING PIG LATIN]. My father and my uncle taught it to me when I was nine. Like me, she's tickled to be able to see things that sound silly, that her mother can't understand. And silliness is important in our family.

Family humor is a very specific thing. Some families like teasing, others like pranks and practical jokes, not mine. Thank goodness. On my father's side, the family loves puns, but on my mother's side, the taste runs to jokes and riddles, preferably absurdist ones. As a young boy, I liked the one of my mother collected, simple ones from joke books and so on.

What's purple and glows? Electric grape. But I think older, my mind craved more twisted, bits of language. I still remember the moment-- I must have been around 10-- when my uncle dropped this gem at the dinner table. Why does a duck? Because one of it's legs are both the same length. Let that giggle around in your brain for a while like a mutant dwerb. One of it's legs are both the same length.

In university, I was floored by the number one item, Letterman's top 10 stupid questions we've been sent, which was, how many raisins in a dollar? So I'm an admirer and a practitioner of the absurd. And so I'm pleased to be in the land that gave us back in the '40s military intelligence and more recently, keep your government hands off my Medicare.

So America has a rich tradition of absurdity, and I assume as a practitioner because as a practitioner of the absurd that I've been asked today to come and talk to you about open source business models. Open source business models? Because we're all an open source for the money. Right? Yeah. Absolutely.

The sense of community, the shared accomplishments, the satisfaction of understanding how something works all the way down, the freedom to build exciting, new things without asking permission first, the chance to work with a global community of like-minded individuals, these are all secondary considerations. The important thing is-- and it's surprising how often this objection to open source is raised. The important thing is how can a traditional software company make money selling open source software?

The logic is straightforward. If existing software companies make money by walling off their intellectual property and sell limited license access to it, it necessarily follows that if a business doesn't not wall off their intellectual property and sell limited license access to it, the business will not survive. QED.

The 451 Group, which is an IT industry analysis company, did a survey of open source companies. And they came up with the best explanation of open source business models that I've seen so far. "Open source is not a business model." So they actually said a little more than that. They said, "Open source is a business tactic, not a business model. Open source is not a market in and of itself, nor is it a vertical segment of the market. Open source is a software development and distribution model that is enabled by a licensing tactic."

OK then, so what's a business model? A business model describes the rationale how an organization creates, delivers, and captures value. A business model is a process that bundles something of value. Where do we create our value?

Business models can be as simple as building item from parts and selling the higher value result, but they don't have to be. The most disruptive business models tend to be the most surprising ones because they convert our notions of where the intrinsic value of a product lies. The initial business model for newspapers and news magazines is very simple. The publisher assemble the newspaper from original writing, laid it out and printed it. Readers paid the publisher for that product. The value was in content delivered to a reader.

In 1729, Benjamin Franklin started publishing the Philadelphia Gazette, and he did something very odd. He sold each copy for a fraction of its production cost, so he was losing money on every copy. But he made up the difference by sales of advertising. He was still commissioning original writing. He was still laying it out. He was still printing it and distributing the product to readers, but the value was in the readership delivered to advertisers, not in the newspaper itself.

The lesson of the newspaper business and many other businesses is that you do not have to sell a thing to capture value with it. You do not have to sell a thing to capture value with it. For a more recent example of indirect value capture, look at the Amazon Kindle 3G. The parts cost, $156. The device cost, $189. That actually leaves $33 to cover all the R&D marketing, cell data charges, anything else anyone does around selling the Kindle 3G. Really, Amazon is giving the Kindle away for free.

But the zero-profit Kindles are enabling a very high-profit business in shipping digital books to customers. So it's not unsurprising if a business invests in an unprofitable secondary business to create value in a profitable primary one. Or consider the venerable concession model. Why does every museum, presumably in the business of showing old stuff and art, have a gift shop? Surely, not an educational purposes.

Why does movie popcorn cost $7.50? Why does an Egg McMuffin cost $2 at McDonald's, but $5 on an airplane? These lucrative side businesses have exclusive access to a buying population that is primed and ready to consume their wares. So there is value in access to a market of pre-qualified customers who are ready to buy. In their study of open source businesses, the 451 Group noted that there were a large number of different tactics for creating value on the open source world and that any given business might employ several of them simultaneously. So here are some of the tactics commonly associated with extracting value around open source. Dual licensing, and this is probably one of the older and most well-known ones popularized by MySQL and Sleepycat Software, both curiously now wards of the Oracle Corporation, offers a GPL version of the software for free, but charges for access to a non-GPL version. The valuable is primarily for companies who want to embed the software for resale, but it also derives from misunderstandings and paranoia around the GPL.

The open core model is practiced by some relatively new venture funded businesses, like Xen and SugarCRM and Nagios. Open core combines an open source core project with proprietary extra features to provide sale value. The sale is in the extra features. Both dual license and open core projects tend to be captive of a single company since in order to carry out their dual licensing scheme, the company has hold the copyrights to the open source project, which usually means they also employ the core contributors directly.

And all these models have in common is that they look an awful lot like a proprietary software companies. Open core, dual licensing, and proprietary licensing all capture value in the same way, by charging for access to intellectual property. And that's perhaps why these models are the easiest to evangelize and receive the lion's share of venture funding over the past several years.

Moving away from the traditional software model, my favorite system is articulated by James Dixon, Pentaho, for what he calls professional open source software. Imagine a bee hive. Now, imagine you want some honey. Knowing nothing about bees or maybe knowing a little bit about bees, you might be understandably leary about just opening up the hive and trying to scoop some out.

What you need is a bee expert, a beekeeper. The beekeeper mediates our relationship with bees. He provides the bees with hives and care, and the bees generate raw honeycomb, which the processes, packages, markets, and sells to us for money, which he uses to buy hives and so on. So even though he doesn't make honey himself, the beekeeper adds value to the product we buy.

In James Dixon's model, open source developers are the bees. How many developers do we have here? Heads up. Developers, would you please buzz. Louder, please. Louder. Users, that's a pretty intimidating sound, right? Maybe you want someone who can help you with the bees to help you get better.

So in this model, the professional open source software company is the beekeeper, and they provide development support to community, and they packed the raw software for use by customers adding things like installer's documentation, uniform branding, messaging, and support agreements. The professional open source company completes the virtuous circle where the value of the raw software produced by the community is fused with value from the company to become a sellable product, which provides resources to employ staff and provides support that in turn strengthens the open source community, providing yet more valuable software, and around and around.

Take apart the proprietary software company and you find development, product management sales, documentation, training. The professional open source company is a result of looking at these parts and asking, is there still value here if we move most of the development and intellectual property outside the organization? And the answer is yes.

Dixon's company, Pentaho, operates on this model, so does Red Hat, So does the company I work form, OpenGeo. Professional marketing materials provide value purchases looking to rationalize their decisions. Integration provides value. Installation provides value. Training and documentation provide value, but perhaps the biggest value sold by professional open source companies is the support contract.

And we say we're providing support, but what we're really providing is insurance for when things go wrong. Real insurance companies protect against disaster by hoarding a large pile of money to fix things afterwards. Professional open source companies protect against disaster by hoarding a large pile of intellectual capital in the form of not disembodied brains, but core developers and their chosen projects.

So Red Hat employees Alan Cox, the Linux kernel guru. And Pentaho employees Matt Casters of the Kettle project. And OpenGeo employs, well, me and many others. So there's a whole other category of open source business that employs the core developers too, so a different business model. This is the product specialist. We have a lot of product specialists here.

The product specialist employs core developers, but they frequently get paid for non-core development tasks. However, because they have core developers on staff, they can add new features for clients. They can perform very complex systems integrations with the open source code. But wait a second. Aren't we just describing consultancy? The goal of the product specialist, though, is to become not just to become a consultancy, but the consultancy. To dominate the mind share for their particular project, to be the biggest fish in their pond.

So Mapgears tries to be the go-to company for MapServer, and Camptocamp tries to be the go-to company for Mapfish. And Oslandia tries to be the go-to company for PostGIS, and RadiantBlue tries to be the go-to company for Awesome, and GeoCat tries to be the go-to company for GeoNetwork to be the biggest fish in their slice.

Product specialists usually spend less money on sales and marketing. They spend theirs spare resources on core development to further hone their reputations as top experts in their fields. Folks back home ask me how do I get invited to speak at all these conferences. And being a modest man, I say, it's because I'm world famous. [INAUDIBLE] the very right small subset of people. The same thing applies to product specialists. The niche they gain to dominate might be quite small, but it's still big enough on a global basis to support a profitable business.

So far, the examples I've looked at, they all look a lot like existing software companies or IG companies. We understand that, but there's lots more valuable value available in the world of open source. Any open source project, if sufficiently popular, probably has a side market for all kinds of things that aren't the software, but add value to the software, that make it prettier. Production-oriented services and add-ons are a form of accessory. So Linux world, the Red Hat Network is an example. JBoss has the JBoss Operations Network.

Deployment platforms. They're an accessory that adds value by making it easy to put open source software stacks into production. So Heroku provides this for Ruby on Rails. PHP Fog does this for PHP. Accessories range from these kinds of highly technical services at one end all the way down to plush toys and CD-ROMs at the other. But just in general, there is value we had in selling accessories to open source projects.

Now, these businesses don't necessarily look like traditional software vendors anymore because they don't necessarily spend much of their time working under core, and they might extract any value directly from the core software. So are these open source companies? Well, they are creating sellable value with their association with open source. So there are all kinds of models for finding value in and around open source, and there are all kinds of open source companies. So Linksys, is that an open source company? They make routers, but they put Linux on it. They derive value through not having to license proprietary OS and through sharing development effort on wireless chip drivers. Any of the handset makers using Android, are these open source companies? They're derived in value through participating in the Android community process.

Which brings me to the largest open source company in the world. It's not Sun. It's not IBM. It's not even Red Hat. It's Google, the big G. Measuring just volume of code released, the 14 million lines only Chrome, Google, GWT, and Android put Google at the head of the pack in terms of global contributions. But it's the fact that Google's core infrastructure, the miles and miles of racks of generic boxes running Linux, that would not have been possible but for open source and for the fact that Google engineers in the communities of open source projects Google depends on participate and contribute. That makes the case for Google as the world's biggest open source company.

But notably the place where Google creates value has nothing to do with open source. Google creates value through information organization. In the words of their mission statement, organizing all world's data, making it universally accessible, and then pasting little ads all around it. But that information organization would not have been possible without open source.

If in the early days Google had to write or license all the code they needed, they would have never gotten off the ground. Without open source, the only organizations capable of running at web scale would have been those who already owned the necessary intellectual property and operating systems and file systems and databases-- Microsoft, Sun, and IBM. Imagine what the web would be runned by Microsoft, Sun, and IBM.

But we didn't get that. Instead we a got a world where success is determined by the ability to quickly gather and manipulate and analyze information, which is why librarians are in such high demand. I like librarians and computer programmers. Marc Andreessen, one of the founders of Netscape and an all-around smarmy bastard, wrote in The Wall Street Journal last month that software is eating the world. And in support of his thesis, he points to numerous success stories, companies leveraging their superiority in software and information analysis in the superiority in the overall marketplace.

Borders being beaten by Blockbuster being put out of business by Netflix. Disney happy to buy Pixar to have animation again. Kodak went under, while Flickr flies up. LinkedIn taking over the recruiting market. The telecom industry being taken down by Skype, even the worldly success of traditional organizations like FedEx and Walmart can all be traced back to superior use of information technology and data management.

These examples show that victory is going to the organizations that can most nimbly extract value from their data streams. And how do we extract value from data streams? With software.

And who creates software? Programmers. And what tools are programmers most effective with? Transparent, flexible, repurposable tools. Tools where the answer is an email or a Google search or IRC question wing. Tools where there are no deployment limitations. Tool where scaling is just a matter of spinning up instances. Tools where the bugs can be caught and squashed in internet time.

Open source tools. Open source tools help you and your developers unlock value in information. Now, it might be intellectually entertaining to look at the models of companies like OpenGeo and compare it to the established proprietary companies and wonder if we're going to make it. Will our open source business model, will it work? Who knows. I think so. But it's not that relevant to you.

You need to have your own open source business model. You need to do the research and understand where open source can add value to your business, to your agency, to your department. And I hope that's why you're all here. So this week, go to the sessions, talk to your developers, talk to your peers, join the open source community, and find your own source business model that's out there waiting for you. Tharbank yarbou much.


The papers and video describe several open source and proprietary GIS systems. For this week's assignment, I'd like you to write down a workflow that you carry out regularly and then discuss how you would carry out this workflow in an open source GIS client (you can choose any you like). Alternatively, you can directly respond to the readings or the video. Post the results in the comments below. To get credit for this assignment, you must post at least one original comment and respond to at least one comment by your fellow students.

Joining the Discussion

The Lesson 6 Reading Discussion in Canvas can be accessed by clicking on Modules in our GEOG 583 space in Canvas. To post a comment, click "Reply" under the discussion prompt and begin typing in the text box, or you can choose to reply to an existing thread. When you are finished typing, click on the "Post Reply" button.