Steven Feldman (@StevenFeldman) is founder of geo consultancy KnowWhere, chairman of geo.me, chairman of Exprodat Consulting, a strategic advisor to Astun Technology, a Special Lecturer at the School of Geography at the University of Nottingham, and chair of the Local Organizing Committee for FOSS4G 2013. He is part of the Taarifa team and helped start the OSM-GB project. Previously he was head of professional services at whereonearth.com and UK Managing Director of MapInfo.
Q: You’re a long-time regular on the geo scene here in London, diving into OpenStreetMap many years ago and sponsoring #geomob the last few years. And yet you’re fairly far from the typical neogeo stereotype. Putting it gently, you’re a bit more experienced than the typical web2.0 code jockey. Indeed geo is actually your second career. What’s your geostory?
A: Let’s get the experience thing out of the way — I just had my Beatles Birthday, you can work that out. I am a commercial animal through and through — I’ve never written a line of code, my biggest technical achievement is tweaking the CSS on my blog.
I graduated from Cambridge with an Economics degree, an idea for a PhD but no funding, and no other idea what I wanted to do. I was offered a job in a mirror manufacturing business, I thought I would take it for a few months while I looked for something interesting to do, and I ended up staying in the building materials industry for over 20 years. I finished up running a division of Pilkington (the glass makers) and then got made redundant at 45. A short stressful and not very successful investment in environmental monitoring tech followed, including a lesson about flogging dead horses which I should share with any startups that I advise. Then I bumped into a friend who owned GDC, a data capture business, that was about to merge with one of those then-exciting internet startups which was about to become whereonearth.com. It sounded like fun and it was a million miles from glass and mirror manufacture so I joined up and headed up the professional services and GIS team at whereonearth. A few years later the whereonearth burn rate was exceeding investors’ patience and we had the opportunity to buy out the old GDC software business which no one thought was sexy enough in the dot com era. We knew that e-government was about to take off in the UK and with some trepidation took the opportunity with all but one of our 18 staff investing their money to buy the business. Less than 5 years later we sold GDC to MapInfo for quite a lot of money and made most of our staff/investors a good bit wealthier. I stayed on for a couple of years as Managing Director of MapInfo UK and headed up product and industry management across EMEA, two years was enough for them and me!
Since 2008 I have been having fun investing and working with startups, doing lots of open stuff because it’s disruptive, advising businesses in the geo industry, and doing a tiny bit at Nottingham University.
Q: When I told you I wanted to interview you for GeoHipster you replied that you’re more of a geohippy than hipster. What’s the difference?
A: I am not sure that I know what a ‘hipster’ is, I hope it is more like James Dean than Henry Winkler. I guess you mean someone who does ‘cool’ or innovative stuff with geo; I don’t think that’s me. I don’t really do anything with geo on the tech front, I am probably too late in my career to start another business even if I had a big idea, but I do know how to build and run a business and I am always up for an investment of time and money in someone else’s great idea. I think I am a reasonable marketeer and evangelist for things I am passionate about (and there are quite a few of those), which can be noisy but isn’t really hip.
I grew up in the sixties listening to Dylan and the Dead, demonstrating against apartheid and the Vietnam war, and believing that our generation could change the world. A first life in building materials grinds some of that idealism out of you, but the last 10 years in geo have rekindled that passion and belief that people can make a difference, particularly with a combination of Geo and Open. Add to that the fact that I banked the ‘fuck you’ money, and I now have the freedom to try and give something back and make a difference — so let me be your first Geohippy interview.
Q: A few years back you were one of the people behind the now defunct OSM-GB project. Tell us about the project and why it’s no longer operating. Was it just too soon? Is OSM the future?
A: That was at the Nottingham Geospatial Institute. We got the funding to use some heavyweight rules-based quality technology from 1Spatial (which is used by Ordnance Survey) to try and build an automated quality improvement process on OSM, and then to explore how OSM might be used by ‘professional users’, particularly in the public sector.
We discovered that we could generate some geometric improvements to the OSM data and we could identify some potential errors both in the geometry and the attribution, but we didn’t want to push our potential corrections back into the master dataset (a lot of what we identified were only potential errors rather than certainties), and we never worked out how to get engagement with the OSM mappers.
We served our ‘corrected’ version of OSM as a WMS and a tile service in OSMGB so that it would be simple for professional GIS users to consume. I was disappointed how little usage we actually got from the public sector despite a lot of initial interest at pretty high levels. The project was funded for about 16 months, we managed to keep it running for a bit longer, but eventually with no one interested in funding us we had to wrap it up.
I love OSM, I think it can be a game changer in some sectors where it is more than good enough. But let’s be honest, in the spaces where I usually work the data is too far from complete, consistent and accurate to be used as authoritative data in most public sector and mission-critical applications. I doubt that will ever change given the producer-centric focus of OSM (we map what we want because we can), but I would love to be proven wrong. OSM, even as it is now, has enormous potential to complement authoritative data from other sources, and we should be continuing to explore how we can make use of it in the public sector.
Q: Relatedly, any thoughts on the recent meltdown of OSMF? Can OSM succeed without a well organized OSMF?
A: Here’s some troll food for you. OSM and OSMF have never really worked out a comfortable relationship. OSMF seems to me to have little or no control or even influence over ‘the map’, its vision, licensing, organisation or strategy. OSMF is split into three camps at the moment:
- Camp 1 wants to keep things ultra-light-touch and leave every decision to the activists amongst the mappers (and probably to not make many decisions, preferring to let everyone ‘do their thing’).
- Camp 2 would like to create a more professional organisation that could raise funding and would provide direction to the project and be able to represent the project to governments and businesses that wanted to engage with OSM (I am definitely in this camp).
- And the majority, even within OSMF, aren’t interested.
The wider OSM community is largely not interested in this stuff and just wants to get on with mapping what they want to map.
The recent meltdown as you describe it is a storm in a teacup with a relatively small number of people shouting at each other in public through the corrosive medium of email lists. You can’t have a conversation on an email list, most people in OSMF don’t even know or care what the argument is about. We talk about a community with over a million contributors, but less than 200 people voted for the new OSMF board; no one cares or understands. So now we have a board which is predominantly Camp 1 and likely to become more so over the coming year with motions for mandatory resignations, etc.
Not the way I would like to have seen things develop, but hey that’s what happens in a ‘community’, and you have to work from where we are. Maybe things will change in the coming years, I would like to see OSM/OSMF realising the vision of becoming the best and the most open map of the world that was used and supported by a colossal number of people and organisations for everyone’s benefit. I don’t think we can do that without fundamental change in the organisation of the project.
Q: Last year you helped organize FOSS4G in Nottingham. For years you’ve been a vocal advocate of open source in geo, and the need for companies to give back to the OS movement (a topic you’re presenting about here at wherecamp.de). As someone with long experience in the industry, tell us your perspective on the rise of open-source and where you see things moving in the future.
A: I am struggling to find the metaphor, “rise of open-source” just doesn’t describe what seems to be an unstoppable torrent or an overwhelmingly inevitable transformation of IT. I am going to confine myself to a short reply on Open Source Geo or we will be here till next year!
Much of what we do with geo today is pretty much ‘known stuff’ — we store data (in vector or raster formats) in a database, we edit it, we catalogue it, we query and render it to the web, mobile or desktop, and that’s most of what we do. That stuff is quite commoditised nowadays and it is inevitable that open source will get wide and growing adoption in those circumstances.
Add to that the fact that most surveys suggest that well over half of GI usage is in the public sector, who are experiencing massive financial pressures around the world and are looking to save costs by reducing their proprietary software inventory.
Oh, and if you want another thought, a lot of users and suppliers are looking to move their geo infrastructure to the cloud to provide a more flexible and scalable solution. Open Source provides a more ‘commercially scalable’ solution because you are not paying a software tax on the success of your application.
Q: You’re an advisor to / investor in several UK geo start-ups. What do you see for the future of the scene? What do you look for in a start-up?
A: That’s simple — people, people, and people. Of course you have to have a good concept and some idea of how that might make money in the future, I sort of take that for granted. I’ve looked at dozens of start-ups and invested in a few, for me it always comes down to people. If the people pitching the concept to me come over as smart, committed, and have integrity, then I get interested (it helps if I like them too). Otherwise just move on, there are plenty of fish in the sea.
I’m a bit cautious and boring as an investor — I want to see some early signs of revenue and a credible business plan. These seem to be quite scarce in the London start-up scene, particular amongst people who have had a great idea involving location.
Q: You blogged once about someone from the corporate world asking why you “waste” your time with small companies. Geo is dominated by giants like Esri, Google, TeleAtlas, Navteq, or national mapping agencies like the Ordnance Survey. Do start-ups have a chance to be globally relevant, or are they consigned to the niches? In your post you conclude small, nimble, OSS companies will eat the lunch of the incumbents. Still feel that way?
A: Hah, I guess that article was bound to come back to haunt me. You can’t consider a dominant software player like Esri (or some of the smaller long term players), a national mapping agency, and a couple of big navigation data providers as if they were the same.
If the big software vendors can’t adapt their business models rapidly they will lose a lot of market share to companies basing their offers on open source, that is already happening in the UK public sector.
I don’t see the mapping equivalent of open source — OpenStreetMap — eating Ordnance Survey’s lunch for a whole host of reasons, e.g. detail, authority, coverage, and consistency. The navigation market is going to come under increasing pressure as OSM moves from ‘good enough’ to pretty darn good, they could find themselves squeezed into high value niches.
Q: Your next challenge is as a non-exec director of the Open Addresses project getting moving here in the UK. This feels like a topic that has been going around forever, I can remember submitting postcodes to the old FreeThePostcode site a decade ago. What’s different now?
A: The Address Wars have been going on for a heck of a long time and we in the open data community are still battling away to get government to recognise that a single comprehensive address dataset is a piece of national information infrastructure that needs to be freely available to everyone for whatever use they may have.
We seemed to have taken steps backwards when the Ordnance Survey mopped up a big chunk of addressing provision by acquiring Intelligent Addressing and the data contributed by all of the Local Authorities, then there was a further setback when the government left the Postal Address File with the privatised Royal Mail. Open Addresses is trying to resolve this long-standing problem by creating a GB address database from a variety of Open Data sources and contributions through crowdsourcing (both bulk contributions and individuals). We think we can get to a fairly usable dataset within a year and have got funding to cover the initial beta phase. Maybe this will be a game changer for addressing in GB?
Q: Any closing thoughts for all the geohipsters (and hippies) out there?
A: You can’t choose to be a geohipster, it seems to be a label that others apply to you if they think that what you have done is in some way cool; I don’t think that is me. I have done pretty regular mainstream things in geo that worked for local and central government, police forces, insurance and oil exploration, that’s probably not geohipster and I’m fine with that.
Geohippies want to make a difference through disruption, geo-evangelism and a bit of altruism (I coined the term so I get to have first try at defining it). Sounds like fun to me.