map butcher

At FOSS4G 2009 this year I had the pleasure of sitting down and chatting with  Arnulf Christl (ac), Paul Ramsey (pr) and my friend Pete Fraser (pf) about Open Source GeoSpatial software and of course the ‘old enemy’. It was my intention to post this interview to geogeek tv however I proved to myself that I’m not the next Stanley Kubrick and so I’ve had to rely on my transcript

Interview 23.10.2009 Sydney, Australia

mb: Is the conference a totally separate beast to OSGeo?

ac: We’ve been discussing this at the board meetings we have. It started off as a separate beast, but now its very tightly integrated with OSGeo. I think it’s very important that it’s a seed process; whenever we go to a place to do a process we leave a seed, which is going to grow. That’s why it is always in a different place each year. OSGeo is a thin layer on top, but it’s the local chapters that bring it in to a local context.

mb: Is having each local chapter set the conference agenda each year a good or a bad thing?

ac: A good thing.

mb: Do each of the local chapters share the same goal?

pr: No they don’t share the same goal. That’s the point. Every region has it’s own problems, piece of software, some regions are very GeoServer focused, some regions are very MapServer focused, some regions are going to be all about the environment, some are going to be all about cadastral infrastructure – which is determined by who the membership is. The only thing we can define is the terms of reference around open source.

ac: In Germany we have a term ‘Leitbild’ (http://dict.leo.org/?search=leitbild) in German which is the big picture. OSGeo is about the big picture, the local chapters bring it into a local context, with their special interests. It tends to broaden our perspective. I don’t think it’s a diluting, but an integrating. We had a big discussion on the discuss list and one guy said ‘don’t let us be sold by the suits’ and other people said ‘we need the suits’ – there’s both perspectives, both are true.

mb: In 2007 there was an amazing buzz around FOSS4G, it’s interesting to reflect this year on where things have moved since then, what’s your perspective on this?

pr: The legitimacy is a lot higher. So the number of people and decision makers who consider it a legitimate option is going up, but, I think my professional experience has been similar to yours in that your feeling is ‘this is going to be huge!’ – ‘why isn’t this huge?’ That’s because it’s slow its generational, you know decision makers who are 55, 60, coming up to retirement, they’re not going to take big chances or risks. We may be old, but we’re just coming into positions where you can make big infrastructure decisions. There will still be folks who’s professional careers were built in organisations who are afraid to do anything without a vendor, but there is going to be more and more people who haven’t and certainly if you look at the younger generation behind us who’s professional teeth have been cut on things like web development where open source is not a mild option, it is the predominant majority it’s going to be a big change.

ac: There’s another aspect, which is special to our world, the spatial world. The spatial world is slower, if you’re going to change your text procssing software you can just change it,… install a new package and that’s it. But if you want to change the software that you use for your spatial data it’s a big thing

*****AC  - talks about coming to a border, and falling off and this is where standards come in – **********

mb: Is there a different feeling now between the open source and vendor worlds? As Dale Lutz said – are we seeing the end of religion?

pr: Vendors are going to have a role, vendors can add value. Proprietary software, if its valuable enough people will buy it. Where that breaks down is when the market reaches consolidation, then you have a vendor just sitting there charging rents, and at that point the dynamic shifts from one in which you have vendors competing to one in which you have a fairly static vendor being eroded by customers who are tired of being charged rents. So instead of them having options brought to them by competing vendors they start to search out options for themselves and open source is so amenable to that because it’s so easy to grab it and try it, and use it.

pf: Is that really the dominant mode that we’re in?

pr: Geospatial, yeh. Databases, to a large extent.

pf: I’m not seeing a great amount of innovation, the zeitgeist seems to be ‘why would you use this proprietary solution, look we’re done this with nothing’. Gordon Moore’s attitude to open source is that it should really be innovating – what do you think?

pr: I think it’s actually wrong. Well at least in as far as if you want to see open source move really fast it has to be innovating. And where you see it moving really fast it has in fact innovated or at least found a temporary crack in the dominant vendors product portfolio. Like we saw MapServer make a big run into ESRI’s web service market when they left the OGC standards stuff to rot for several years. Agencies wanted to do it, and open source leaked into the crack really fast.

pf: So it wasn’t performance and cost it was the open standards?

pr: Yeh it was in fact being able to support that new feature that the vendor had decided strategically it wasn’t worth it for them. So they lost out. You almost saw it happen in the 2005 – 2009 period with the upending of the old web mapping paradigm and its replacement by the GeoWeb 2.0, because again it looked like ESRI was just going to hold back the tide – it appears they just got very lucky picking up some hires who understood and bring that vision to them, but they weren’t generating it themselves. They almost lost it because tools like MapServer were coming out with Restful resources, OpenLayers was pushing that paradigm far far faster – again they had to close that hole. The vendor often moves slower.

pf: What would you do if you were running ESRI?

pr: Oh, I’d do exactly what they’re doing. There’s nothing wrong with what they’re doing, they’re doing a great job, and they want to defend their market. They operate primarily a defensive strategy, they’ve got lots of room to do it because their customers move relatively slowly. Only a narrow leading edge says ‘that new thing looks great I want to do that’ and leaps ahead of the vendor. For the most part the customers are happy to be lead.

ac: I think the open source movement is an evolution not a revolution, we’re in a transition phase, perhaps 5-10 years eventually proprietary solutions will occupy the niche. Paul’s presentation was just perfect the position of the money switches. It’s good for the local economy too. In Germany we have a lot of small shops, which use Open Source, and the money stays within the country.

mb: Do open source geospatial projects re-invent the wheel over and over again, how is re-use organised and controlled?

ac: It’s not controlled, it’s complete chaos. I was sitting in a presentation just then thinking shit I did something like this 5 years ago and no one took notice – our marketing sucks! We didn’t get the message across. So these guys didn’t know about our project, but now I’ve spoken to them and at the code sprint we’ll sit down and see what they have and see what we have and see if it can be put together and ideally make it better. But basically its chaos and to manage chaos you have to communicate. That’s the point of this conference, this meeting of the tribes where all the people get together.

pr: The core value of open source is not licensing or source code, its communication. When communication is maximised that’s when the best outcomes happen. I had the same thing happen with the Udig project back in 2005 – I carefully went out, wrote my proposals, got my grant money, worked on it for 6 months and then learned on a well funded project in Spain basically doing the same thing. Because neither of us broadcast our intent widely enough.

pf: That’s fine that’s parallelism.

pr: Its parallelism but, particularly when you’re looking at funding buckets and big projects that you try to get up and over the hump of utility at the start being able to work together would have been invaluable, although perhaps not possible. The other thing about big funding buckets is they tend to force you to deadlines, which makes working together more difficult

ac: I think what you said is perfectly right – there’s no problem doing things twice, or 3, 4 5 times. I think where OSGeo has a role is at the global scale where we can help to coordinate thinking once, coding once and using several times over. That makes a lot of sense. As you start to deploy this software around the world then you see the growth in communication start. So maybe you wasted some time and money doing things twice, but you’ve learned a great deal from both projects, and whatever comes out of it is hopefully better.

pf: What are the new risks, threats and new lock in?

pr: The new lock in is data obviously and that battle is already drawn, it’s amazing. We already have open street map up there as the open source alternative in data. We already have proprietary silos of even user-generated data being built, which is scary. Some people opt for freedom and some people opt for semi-freedom. The irony is that all these infrastructures are being built on open source software so even then they’re building closed data they’re usually building it on open source software. The cloud can only run on open source because you can’t afford to scale when you’re paying per license. No one but Microsoft is building Microsoft’s cloud.

pf: Do you think Stallman is going to look out the window one day and think oops it wasn’t about software at all? Software’s not the story – the big story is..

pr: Step by step. There was no data story until there was ubiquitous connectivity. Of course it was ubiquitous connectivity that allowed open source to happen in the first place. Without the Internet there’s no open source. It’s amazing actually that Stallman started doing open source before the widespread availability of the Internet. That makes it really hard to find collaborators.

ac: It couldn’t work – it’s the same thing with OSM when it popped up everybody said this wouldn’t work, same with wikipedia. Actually wikipedia doesn’t work…I tried to upload the OSGeo logo to the Spanish wikipedia up and wasn’t allowed because we don’t have a copyright on the media so I ended up in wiki commons. There are so many restrictions, legal entities and complications. I see this as one of the greatest problems for the open source movement.

pr: That’s actually a place where there is a serious asymmetrical situation around interpretation of rules. What are the rules and how should we interpret them? Are we going to get in trouble if we break them? As little guys we can’t afford to screw up too much. One of the most interesting comments I saw after Google released its new layer of parcels on a big long thread of James Fees blog someone said ‘I work for that council, and those are the parcels that I checked and we don’t give out our parcels for reuse like that – how did it get there?’ and the response was ‘Google’s big enough they can just litigate it out’. I’m glad Google’s breaking the rules because that’ll force uniformity of public access across all these counties who were previously, in a heterogeneous way enforcing rules in very different ways for different people. But it does show Google to be the big gorilla again, they change the rules and the rest of us have to play by those rules, so you end up with wikipedia where you can’t upload you’re own damm logo, but meanwhile Google can scrape it and put it on Google images with no trouble at all. The 600lb Gorilla gets to decide what the rules are going to be. Increasingly Google is dictating the rules around data.

pf: There was an interesting post on the New Zealand ESRI User group list this year where some guy had got a million bucks worth of brand new imagery and it was beautiful and he posted saying, ‘I’ve cracked Google, I’ve got my data onto Google and this is how I did it’. This was an achievement for him to have given a million bucks worth of data to Google. Bless him he had the right idea he’s sharing.

mb: Is what makes FOSS4G different the fact that it’s a ‘working’ conference? We’re sitting here in the Sydney Convention Centre we seem a long way from the ‘un-conference’?

pr: Because we’re serving this dual purpose, on the one hand its getting the tribes together for a mass geek-out, but on the other hand a large proportion of people at this conference are just dipping their toes in the water – for them its very much a learning experience. The ‘un-conference’ probably wouldn’t be the nicest environment.

ac: I think what I’ve seen in this conference is that it’s a perfect mix. I don’t know why it always works, but it does? Maybe its because there are so many voices shaping the conference. In Germany the FossGIS conference is developed entirely through the wiki, unless you get bigger, like four or five hundred people then it works. So FOSS4G for me has always been a big thing.

mb: Is it a challenge then to try and gather the impressions of the conference from those who are just dipping their feet in?

pr: It would be great if we could get 100 words from everyone who attends! The reality is that geeks control the volume on the Internet. Most of our sources of news are geek controlled. One of the great things about 2007 was having Edina there reporting and summarising, because at the end someone who works in the general geospatial marketplace was able to look at the conference from that perspective.

ac: We’ve got to work on that a bit. I think we have to connect more with standard media to get better coverage.

mb: If the whole process of open source maturity is a slow burn, OSGeo can afford to try a number of different governance approaches?

pr: People are generational with technology too, the software gets old people change. You build a system, run it for 15 years then re-evaluate, how you’re going to build it next time. And each of those decisions you come up with a different answer right. Thirty years ago you built your system on an IBM mainframe, 15 years ago you built your system around Oracle. Now…

ac: We already have this in a slightly different context, when open street map came into existence, they were like despising us, OSGeo guys, because we do the standards. They mixed it up – they never got the difference between OSGeo and OGC. Now if you go to a cadastral agency in Germany and you ask the difference between OSGeo and OGC they know, but the open street map guys just dumped us in the same place, because they thought, they are the young guys, they are the revolutionaries.

pf: I think they are. Listen to this conversation, you’re slipping into this old worldly industry thing, ‘it’s evolutionary, not revolutionary’. That whole model of the economy is endangered. You’re already playing an old game.

ac: ..and they will. Just let them stew for 3 or 4 more years, they will do the same thing – we’re getting older

pr: they’ll stop going out mapping after beers, they’ll head home to their families.

pf: ..seduced by power

pr: ..some already have.

mb: Pete, as a first time attendee of FOSS4G what are you going to take away from the conference?

pf: I just don’t have enough bandwidth for the conference. I want to spread it over a month. The best bit for me was when someone said to the guys doing MapServer ‘ why would I pick MapServer over GeoServer?’ and the guy gave a little speech on how good GeoServer is! It was fantastic. If you’re going to get straight talk like that –  it’s about the people right.

mb: There’s been a lot of talk at this conference about the right business model for open source, Paul you set the tone in your plenary

pr: It’s not unique to Geospatial, the open source world is still casting around for the right business model, a lot of that is because we’re in this transition period, so there isn’t a ‘right’ business model right now. Again read Matt Asay everyday – he always has something insightful to say about how open source is meshing with the marketplace. One of blogs in that sphere did a survey of venture capital funding to open source companies and how that money differed by model. So prior to the 2000 meltdown when everyone was ‘Linux is going to rule the world’, next Sunday piles of money was being put into 100 percent service based open source start ups, and after the meltdown that peered off. What came back was a pile of funding around ‘open-core’ open source business models where there is a core open source piece, but the things which is developed by the company is that core plus a shell of proprietary stuff, which presumably makes the product more interesting to the eventual buyer, and that trend went up for a while then it came back down and we’re seeing more service based models appear again. No one knows yet, but its clear that the value has to reside out of the software, and whatever value is there has to be compelling enough to get the CTO to take his cheque book out.

ac: What I notice is that there is a transition in the traditional shops that have been around for 10-15 years and they are beginning to turn and offer open source solutions as well as their traditional offerings. I think the same is even happening with the big shops. ESRI is very active in the open source community, like very undercover, but they are involved because they see it is the future and they’re not dumb.

pr: It’s a hard change for them to make, because if your dominant revenue stream is software licenses its very difficult.

ac: If you have 4000 people and 3,500 of them sell boxes and now you want to switch to a service model it just takes years to educate all those box sellers to be able to deliver services.

pf: I think in some ways Jack wouldn’t mind if people did do the easy things and came to ESRI because they were about solving the hard problems. That’s why I asked what would you do if you were running ESRI?

pr: If I was running ESRI I would maximise revenues for as long as possible, then I’d shut the company down. From a financial position that’s probably the right thing to do, its like a forest – the optimal harvest strategy is to raise it flat and then go invest your money in movies. There is a lot of room within that company for different visions other than maximising revenues.

pf: He likes gardening apparently – I think the open source world would appeal to him. I tried to employ him once, I said you want to come and work for me, he was interested, ..he bit, but I’m playing the long game. I’m going to hire Jack.