WEBVTT

1
00:00:06.000 --> 00:00:15.000
<v Speaker A>Welcome to the GeoMob podcast where we discuss geo innovation in any and all forms, whether for fun or profit.

2
00:00:15.000 --> 00:01:07.000
<v Speaker B>Hi everyone, Jeremy here. Welcome back to the GeoMob podcast and particularly the Geospatial Origins strand that I've been leading. I'm very pleased today to welcome Peter Batty to the podcast and this strand. Peter has long histories we'll get into in the geospatial sphere, both in terms of desktop, sorry desktop, geospatial systems and innovation in that sphere, and then later on in the kind of telecoms and mobile sphere as well. And we'll touch also on kind of the rise of open source alternatives and where that's led us as well. So. Hello Peter, welcome to the podcast.

3
00:01:07.000 --> 00:01:10.000
<v Speaker C>Hi Jeremy, thanks for having me.

4
00:01:10.000 --> 00:01:42.000
<v Speaker B>And I think we were looking back to our kind of first contact points and although while I was running the MSC at UCL we had a couple of small world licenses when you worked for that company, I don't think we met at the time and I think it was through the kind of conference circuit of things like the AGI community conference and so on and then in the open source communities that we first met. So kind of 2000s early 2010s, I think that's about right.

5
00:01:42.000 --> 00:01:57.000
<v Speaker C>Yeah, I think so. I was looking back and think the 2009 AGI Geo community, we had some interaction doing something there and then the phosp 4G a few years later. So yeah, it was on the conference circuit around those times, I think.

6
00:01:57.000 --> 00:02:18.000
<v Speaker B>Yes, interesting times. Anyway, let's maybe scroll back to the start of your story. Everyone has a different tale about how they got into what now is broadly called geospatial. Do you want to go back to your start and say how you ended up in this domain?

7
00:02:18.000 --> 00:02:54.000
<v Speaker C>Yes, sure. It was quite accidental and I know quite a few people for who that's the case, but I studied maths at college at Oxford and as an undergraduate and then did an MSc in computation, as they called it, Computer Science. And I joined IBM from there and at the time they were sort of the dominant computer company in the world, both hardware and software. Kind of like going to Google today or well, maybe Google is outdated now, anthropic or somebody. But you know, it was sort of the big tech company of the day

8
00:02:54.000 --> 00:03:02.000
<v Speaker B>and they had both the reputation of the salesman in blue suits and also the motto that no one ever got fired for buying IBM as well.

9
00:03:02.000 --> 00:04:37.000
<v Speaker C>Yes, indeed. And we did actually largely wear blue suits. I did wear suits to work back in those days, which I haven't Done for a long time. Yes, so yes, that's right. Well, maybe further in this story we'll talk about when they might have got fired for buying IBM. But yes, those were both correct statements. But yeah, so I joined IBM just as a sort of general systems engineer, which was the role was kind of a mix of technical and customer facing sales kind of stuff, I guess. So I thought it would give me a good background in different things. But on the first day they told me that there was this new product they brought from the US called GFIs and they wanted me to spend time reading the manuals and learning it. And the end of the first week I was kind of the UK expert as I was the only one who'd really cut it. And I ended up doing the same thing for quite a while. And GPHES stood for GEO Facilities Information Systems. So it was really utility focused GIS product. And it ran on IBM mainframes which typically cost millions of dollars or maybe large number of hundreds of thousands for a small one. And the graphical workstations we used were I think originally around $100,000 and maybe down to a few tens of thousands of dollars over my time there. But so it was definitely.

10
00:04:37.000 --> 00:04:39.000
<v Speaker B>What sort of period was this?

11
00:04:39.000 --> 00:06:16.000
<v Speaker C>So this was 1986. One of the things that's worth highlighting technically is that IBM was really the first company who built a GIS that stored all of its data in a single standard database management system. And that was something I got quite involved in advocating for, but at that time that nobody else really did that typically for performance reasons. Largely the data was split between sort of one storage mechanism for the graphic or geospatial data and another system for the tabular data. So ARC Info got its name. The ARC was the portion that stored the spatial or graphical data. And info was a simple tabular database. But IBM sort of stored everything initially in its hierarchical database DL1 and then moved to its relational database DB2, which was the main focus when I was there. And I got quite involved in advocating for that. Why that was a good thing. There were some challenges with IT performance and so on. But I wrote a paper in 1990 called Exploiting Relational Database and Technology, which ended up being somewhat influential. I think Oracle cited it when they launched Oracle Spatial in 1995. I think it was. There's an influence on them and I still get notifications from this online site, I think it's called Academia Edu and they're probably just trying to get me to sign up, but they keep telling me paper or Quoted it. So it still has a life. But yeah.

12
00:06:16.000 --> 00:06:35.000
<v Speaker B>Did you ever find at IBM, I mean, in terms of the IBM story, you came in part, part way through in some sense from picking up gphist. Did you ever find the. Or hear from the designer there and why they chose that, that method in the first place?

13
00:06:35.000 --> 00:07:14.000
<v Speaker C>I spoke to some of the early people. I mean, I think, you know, the main things that I advocated in that paper were related to that of, you know, that they came from a database background and generally had this concept of, hey, you know, you make a translation, a transaction, updating a feature, a pole or a transformer or whatever it might be in your electric database. You want everything to fail or roll back and it should be integrated. You want one mechanism for backing it up. If you have this sort of tightly related data in two separate systems, there's a lot of reasons why from a database technology perspective that's a bad thing.

14
00:07:14.000 --> 00:07:19.000
<v Speaker B>But synchronization and backup and all and rollback and all those things become more difficult, don't they? Yeah.

15
00:07:19.000 --> 00:07:44.000
<v Speaker C>But then, you know, the flip side of that was, especially back then, it was a performance challenge and that was why people took the route of storing graphical data in some more proprietary formats. So, you know, there were challenges with both, but that became the dominant way of doing things, although, as we'll come and talk about, but my angle, and it changed slightly when I moved to small world. So we'll come back to that.

16
00:07:44.000 --> 00:07:52.000
<v Speaker B>But there's a step in between, isn't there, of IBM trying to develop something a bit. Was it something more general than GFIST they were after in GIS terms?

17
00:07:52.000 --> 00:08:43.000
<v Speaker C>Yeah. So an interesting thing that not many people know about, I don't think, was that IBM decided shortly after I joined, well, a couple of years in, I guess, that they were going to develop a major new GIS product. That GFIS was pretty successful, but there were a number of areas where we felt we could improve it and broaden its focus. And so they decided they were going to develop the world's GIS product, the best GIS product. And it was a bit of a career break for me because I was pretty young, I was 25 or whatever out of college, and they chose me to be on the team that worked on the sort of design of that product with a lot of more senior people. But anyway, it became a huge project. I think at the peak they had 200 people working on it, invested.

18
00:08:43.000 --> 00:08:46.000
<v Speaker B>Very significant. Yes.

19
00:08:46.000 --> 00:09:47.000
<v Speaker C>But, you know, long story short, we could talk about it for a while, but it Basically ended up failing. And I think that ended up indirectly or well, maybe somewhat directly being a factor in Small World success because the IBM had told all their customer base that gist was coming to an end and they were going to get this great new product and the customer base was all ready to move. I actually left shortly before they officially pulled the plug on it. And I'd also been working on how to migrate from the old system to the new system which had some quite strong differences. So I knew the old GFIS system structure very well and then I went and joined Small World, which we'll talk about shortly. But because the GFIS market were all kind of in the mindset of they needed something new, I think we were quite successful, those customers. And if that hadn't happened, you know, Small World might not have grown as quickly, I think.

20
00:09:47.000 --> 00:09:55.000
<v Speaker B>What were the factors in that product failing at IBM? Was it technical or more to do with marketing business?

21
00:09:55.000 --> 00:11:00.000
<v Speaker C>It was technical and you know, I think it's, I'm certainly a big advocate of to do something disruptive or for development in general. It makes a lot of sense to have relatively small teams and also to prototype rapidly and iterate. But this was a more classic waterfall kind of development where they had three tiers, a database, a middle tier and a front end and big spec documents on how they would talk to each other. And so they were developing these big teams and then quite far into the process it became apparent there was some sort of mismatches or missed expectations. And the interesting thing was Small World started almost exactly the same time and put six people on their product and Smallhold wrote a programming language from scratch and a database management system from scratch as well as a GIS with six people in the same time. And they ended up with a revolutionary successful product and IBM put 200 on it and didn't get there. So I think there's a lesson there.

22
00:11:00.000 --> 00:11:08.000
<v Speaker B>Yes, there's some interesting efficiency curve out there is that, that I've not seen about kind of project size versus staff size on the project.

23
00:11:08.000 --> 00:11:09.000
<v Speaker C>Yeah, yeah.

24
00:11:09.000 --> 00:11:19.000
<v Speaker B>So, so what triggered your move from IBM then if it was before that project was completely canned? What, what, what prompted you to move?

25
00:11:19.000 --> 00:13:01.000
<v Speaker C>Well, so I, I'd been in the US for the winter, this was probably when was it? 91 to 92 I guess. And I'd been there working specifically on the migration capability from the old product to the new product. And so I'd had quite a good view of what was going on and I really had concerns about whether it was Actually going to work. And in the meantime I'd met a couple of the guys from Small World. One guy in particular, Mike Bandoc, was actually giving a talk, I think at ucl, the British Computer Society. He gave a talk about a small world that I happened to see and it sounded kind of intriguing. And I had a chat with him afterwards and he was saying, oh, you should come work for us. And I said, well, you know, there's a big project going on, let me see how that goes. But when I got back from three months in the States, I went to talk to the Small World guys and I was just super impressed that while IBM had been trying to look at the best features of all the products out there and put them all into one thing, Small World had several revolutionary new ideas that just IBM wasn't even thinking about and nobody else was. And so I really felt that that was the future. And then I left and people knew I had something of an inside track on the IBM thing, but they hadn't announced anything. And I went from IBM, which was 300,000 people, to this small company in Cambridge, which was about 30 people at the time. So my phone was ringing a lot to say what do you know about what's going on? And I had to, well, not say too much, but it ended up being a really good change.

26
00:13:01.000 --> 00:13:17.000
<v Speaker B>And how did you find that culture shift? IBM, a very American company as well as very big and spread out company, lots of divisions to Cambridge and that smaller, effectively startup company.

27
00:13:17.000 --> 00:14:29.000
<v Speaker C>Yeah, I mean it was a big change, but yeah, overall I loved it. And I've alternated between small and large companies and definitely lean towards the small end of the thing. I mean, both have pros and cons, but it was a fantastic team. They'd worked together on a couple of previous products and just a lot of really top people in different technical areas as well as people good on the business side and the marketing side. And they genuinely had some really significant innovations. So, you know, looking back, you know, you don't realize at the time necessarily what a unique situation you're in, but I would say definitely Small world in the 1990s was the most enjoyable part of my career where we had this product that was way ahead of the market. I came over to the US after a year and we just grew like crazy in the US against established players like ESRI who weren't as dominant. They were still probably the biggest player, but not so much in utilities, but they weren't nearly as dominant as they were now. But we consistently beat ESRI for the first five years over there and Intergraph who were the other main established player.

28
00:14:29.000 --> 00:14:57.000
<v Speaker B>I think that's why when we were running the MSC and we wanted to keep some other giss on board for the students to have experience with apart from esri. I think that's why Small World was in there is because it felt like quite a different beast and revolutionary in its features. Do you want to say a bit about how you think it was kind of different in technology?

29
00:14:57.000 --> 00:16:31.000
<v Speaker C>Yeah, I can briefly touch on it a little conscious of time as well and I could talk about a lot, but. And I did write a blog post which I think we can link to if we post some things. But basically very briefly, they had a very revolutionary database technology and it was interesting because I'd been advocating for storing everything in a standard database like DB2. Small order developed their own specialized relational database that handled long transactions very well, which are a key thing in utilities. And I think we won't get into that now, but about handling the design and construction process, they had a very new way of handling that and then the architecture of that led to really amazing performance as well, where you could just pan and zoom interactively over the whole database, which you couldn't do with other systems at the time. And they had extremely flexible customization environments as well. A language called Small World Magic, which is very Python, like ahead of Python. And then maybe the last thing, which might be something we come back to, but very good network modeling, which is a key thing in utilities and telecoms that is sort of unique to those markets of handling the way that those networks connect and how signals flow along them or gas or electric or telecom signals go. And that was always a strength of Small World. And I think a relative weakness of ESRI that ESRI had good broader functionality, but networks have traditionally been a sort of weakness of them relative to some of the other companies.

30
00:16:31.000 --> 00:16:44.000
<v Speaker B>Yes, yes. It felt like with ESRI at the time you could import your data and then build a network and then build something on top of it. But it is a very multi stage thing rather than being as inherent.

31
00:16:44.000 --> 00:16:46.000
<v Speaker C>Yeah, yeah.

32
00:16:46.000 --> 00:16:58.000
<v Speaker B>And as I recall from the time as well, it was a more kind of object oriented sense in the database as well, as opposed to lots of pieces of data that connected together.

33
00:16:58.000 --> 00:17:51.000
<v Speaker C>Yeah, exactly. It was. So the magic language I mentioned was object oriented and that was again a very new thing at that time, very powerful. And the integration of the database was very strong. Another interesting technical feature was rather than characterizing features as being points, lines or polygons which most traditional GIS systems did and still do. Smallworld had the concept of you had a feature, but it could have multiple spatial attributes that were points, lines or polygons. So you could have more complex objects that may have a point geometry and a polygon geometry, say or two point geometries that interacted in. And this was good for modeling complex equipment and utilities. And so that. That was an interesting thing that I haven't really seen another company do in general. So.

34
00:17:51.000 --> 00:17:57.000
<v Speaker B>Yes. And this was a PC architecture as well by then, wasn't it? Not. Not kind of mainframes and so on.

35
00:17:57.000 --> 00:18:34.000
<v Speaker C>Yeah, well, primarily Unix workstations initially actually in when I started there and then when Windows NT came out, it. We moved to Windows NT and belatedly sort of had something of a web architecture, but it was primarily a desktop client server architecture and had some very clever caching which made for good performance. So the database was. There was less work happening on the database server which also made it scalable to very large numbers of users. There were a lot of kind of intriguing things about it. Technically.

36
00:18:34.000 --> 00:18:55.000
<v Speaker B>Yes, I think anything in the 90s like that is worth a bit of discussion about that hardware because that kind of technology environment was moving so quickly at that time. I think that was almost one of the challenges as much as anything, as much as your software design, it was your hardware target. Felt like quite a moving thing at the time as well.

37
00:18:55.000 --> 00:19:39.000
<v Speaker C>Yeah, exactly. And it's also amazing thinking back when if you look at Moore's Law, you see slightly different variants of it, but it says, you know, performance doubles every 18 months. Say if you take that number, then that means in 15 years performance increases by a factor of 1,000. So if you're going back 30 years, that's a factor of a million. And even if you tweak that a little bit, you know, it's kind of shocking how much we were able to do at today's system in some cases are not even as fast. So yeah, it's quite interesting looking back at what we were able to do with quite limited hardware compared to today.

38
00:19:39.000 --> 00:19:41.000
<v Speaker B>Yes, you had to be much more efficient in that resource, didn't you?

39
00:19:41.000 --> 00:19:43.000
<v Speaker C>Yeah.

40
00:19:43.000 --> 00:19:53.000
<v Speaker B>And so I think particularly there was some specialization here. The Power on system, Is that one of them?

41
00:19:53.000 --> 00:21:43.000
<v Speaker C>Yeah. So one other interesting thing that I worked on was an application called Power On. So I led the development of that from the beginning and it was focused on electric outage management, which is the process of when the power goes out, basically managing the work, figuring out who's out of power, what's the likely problem and managing the process for restoring it. So it's quite a real time, mission critical type of application. And in general the main GIS platforms I think up to that point hadn't been used for really mission critical things like that. There was a separate category of applications that would do that that were very specialized and didn't have such good mapping graphics capabilities. And so we were looking to take the more detailed graphics and network modeling of GIS into that environment. And Smallworld was strong in that electric utility market already, but we felt we could develop this as a separate application on top of Small World. And so that ended up being an interesting project. We did actually use Oracle as the database for that because the Small World database was very specifically designed for these long transactions, but for outage you need short transactions because it's all about real time state of the network. And I came up with a sort of clever way of storing network model for radial network model for radial networks in Oracle, which was kind of a key thing to enable us to do that because in general doing network tracing through a relational database is slow. But we had some good success with that. We ended up becoming the market leader in the management space within a couple of years and at our peak I'd say we probably had 100 million odd people whose electric supply was made more reliable with power on. So that was quite satisfying.

42
00:21:43.000 --> 00:21:44.000
<v Speaker B>That's quite a number, isn't it?

43
00:21:44.000 --> 00:21:48.000
<v Speaker C>Yes, yeah,

44
00:21:48.000 --> 00:21:49.000
<v Speaker B>yeah, sorry, finish off please.

45
00:21:49.000 --> 00:22:18.000
<v Speaker C>Oh, I was just going to say randomly I had a visit. I visited one of the larger utilities in Brazil sometime later on, I think maybe like 2018 or something. I was actually on a mission for the Japanese government, which is another story, but we went into this utility and saw their control room and they were running power on, you know, 20 odd years later. And I knew there were some customers in Brazil, but that was kind of cool to see that it was still

46
00:22:18.000 --> 00:22:26.000
<v Speaker B>going there and was that still being maintained and so on, or was that some old copy that was out?

47
00:22:26.000 --> 00:23:34.000
<v Speaker C>Yeah, so GE has continued to maintain it and there were still quite a number of customers out there. And sort of the same with Smallworld. I mean, maybe Smallworld more generally. GE's continued to maintain it. They hadn't done any real radical RE architecture of it, but I think Small World is still, I would say, running quite a lot of the largest and most complex GIS systems and utilities and telecom. I had a quick look and I think not 100% sure, but I think probably half of the top 10 electric utilities in the US are still using it as their main system of record. And Deutsche Telekom had one of the biggest, most complicated GIS systems with all sorts of incredibly complex integrations. So it's very difficult to move off. But I believe they're still using it. So, you know, it's had a really good run, I mean, having started in the early 90s and to be still going now, 30 plus years later is pretty impressive when it's all those changes in the industry.

48
00:23:34.000 --> 00:23:43.000
<v Speaker B>Yes, but at the same time solves the basic problem, doesn't it? The networks haven't maybe until possibly very recently changed that much in their detail.

49
00:23:43.000 --> 00:24:19.000
<v Speaker C>Yeah. And at the same time, you know, that there has been in recent years a reasonably steady, you know, slow but steady migration towards esri, who are the main kind of active player left in the space, both Small World and Intergraph who kind of didn't renew their products and both have these sort of 30 year old products. Esri has gone through iterations of improving their products in the area and so they're kind of somewhat taking over that market though we can maybe talk a bit more about that in another section.

50
00:24:19.000 --> 00:24:41.000
<v Speaker B>Yes, and just going back to kind of Power on then in its kind of original version at the kind of end of the 90s, early 2000s, that was connected to the live network state, was it? So, I mean, in some ways it feels like what we would now call a digital twin system as much as gis. If it was connected to the control systems.

51
00:24:41.000 --> 00:25:41.000
<v Speaker C>Yes, indeed. So, yeah, it could be integrated with sensors, although it was surprising to me in a lot of cases how little the power company knew about the state of the network. And actually the primary input was customers calling in to say I'm out of power. That gradually evolved as they've had smarter networks. You hear about the smart grid. But when we started there would be a few smart devices on the network, but the main source was just customers would call in. So a big part of what Power on did was it would look at the pattern of customer calls and if there were one customer called in and nobody else in the area, it was probably, probably an issue specific to that customer's connection. But if several called in who were all fed from the same transformer, you suspected it was the transformer. So we were kind of getting near real time information, but then making informed guesses as to where the problem was to direct the people to the, you know, the workers to fix things.

52
00:25:41.000 --> 00:25:43.000
<v Speaker B>Times have changed a bit, I guess.

53
00:25:43.000 --> 00:25:46.000
<v Speaker C>Yeah,

54
00:25:46.000 --> 00:26:03.000
<v Speaker B>yeah. So you mentioned GE already as Part of this picture, General Electrics. What came about then for the sale of Small World, how did that come about and what was your experience through that?

55
00:26:03.000 --> 00:28:26.000
<v Speaker C>Yeah, so Small World got bought by GE in 2000, you know, which was a great endorsement of what we'd done, that GE was huge in the electric industry. And I think we were all hopeful that that would give us more funds to really invest and take over the market. And GE had very strong relationships with the electric utilities and, you know, and all those things ended up being the case. And I think GE partly bought it because GE Power Systems was a very successful business and GE had quite a lot of different businesses that were somewhat independent. And I think GE Power Systems was very strong in the hardware business of selling transformers and equipment for substations, that kind of thing. But they were, were somewhat old school and they were looking to get more into the software side of the business as well as the hardware. So they acquired Small World and a few other companies. But I think in a way the way that the business work was a bit old school and we were all kind of entrepreneurial. And so, as sometimes happens in these cases, at least for some of us, there was a bit of a culture clash. But in general the acquisition worked reasonably well. And so Small World is still being supported out there 20 plus years after. 26 years, I guess, after the acquisition. But a lot of the early employees, myself included, felt it was time to go and do something else. And so, you know, I moved on with three of the other members of the management team to start a company called 10 Sales. I think 18 months after the acquisition, which was the minimum time we had to stay according to the terms of the deal. So we went to new pastures. But, you know, GE was a safe home for it, I guess. But I think at that time we kind of hoped that Smallville could become the dominant GIs in the world. But you know, in the end it's had a good life, but it sort of faded away. And ESRI has a pretty dominant position in the utility market these days.

56
00:28:26.000 --> 00:28:37.000
<v Speaker B>Yes, it's a shame. It's the difference between someone buying a company to just cover their position and improve their position in the market rather than really understanding what they've got and can deal with, I suppose.

57
00:28:37.000 --> 00:29:39.000
<v Speaker C>Yeah, yeah. And to some extent, I think, you know, at the time they bought Small World, you know, it was still a very clear technical market leader, but it was also to some degree approaching a point where you might need to think about re architecting things, though there wasn't perhaps Wasn't an obvious, there weren't really obvious things to do at the time. But in some ways ten years later we built what I think could have been a successor to Small World iqgo Ubiso, which is jumping ahead a bit in the story. But, but anyway they never really invested in sort of creating a new architecture. And I think at some point along the way, you know, it's a credit to it that this last 30 lasted 30 years. But at some point you need to take a somewhat clean slate and say we'll re implement a lot of the good ideas in a new technology architecture, whatever. And that did happen.

58
00:29:39.000 --> 00:29:53.000
<v Speaker B>So after leaving then it was a period of kind of quite a lot of changes for you. Do you want to maybe paint that picture through to the kind of UV sense time?

59
00:29:53.000 --> 00:31:17.000
<v Speaker C>Yeah. So again trying to go fairly quickly but when we started, so we started a company called Tensales, the four of us and loose idea was, and we had myself who was the cto, we had a CEO sales leader type person, Richard Green, David Terrio was sort of had a lot of experience in product management, marketing and Martin Cartwright was the cfo. And so we felt we had kind of a management team that could help more technical founders grow their businesses. So these days people talk about fractional CEOs, whatever it was, that kind of thing, but people weren't using those terms then. And so we got involved with a few companies in Cambridge and long story short, the first one that or what ended up being the primary focus was a company called Ubisense that was a spin out from Cambridge University. Four founders from there and they had some really smart technology for indoor location tracking. And so we worked with them for quite a while and that was a really interesting business. I probably won't get into that again in the interest of time, but it's, it's still intriguing to me that that's still largely an unsolved problem. It was interesting how difficult it was and still today it's really not a solved problem.

60
00:31:17.000 --> 00:31:21.000
<v Speaker B>What was their approach at the time then? This was early 2000.

61
00:31:21.000 --> 00:32:40.000
<v Speaker C>Yeah. So they used technology called Ultra Wideband and that had certain characteristics. So it's a radio frequency type technology. So it was both a hardware and a software company. They had sensors and little tags that you could buy and then accompanying real time software system. But there's a lot of challenges with getting reflections off walls and signals being blocked and that kind of thing. And they were really focused on the high precision end of the market and ended up doing quite well in manufacturing like Car manufacturing, aerospace manufacturing. There were applications in healthcare for tracking machines in hospitals, transportation for tracking buses in depots and things like that. So it was very interesting and sort of broad space and I think the company did reasonably well, but sort of struggled to break out from being a pretty high end, fairly technical thing into more of a mass market thing that was easy to implement. And there's various kind of lower end technologies for indoor location tracking. But I think it's still certainly true to say that there's a lot of unsolved problems in that space still, and it's more complex than you might imagine.

62
00:32:40.000 --> 00:32:52.000
<v Speaker B>Yes, and I think you've kind of alluded to this as well. It depends what your purpose is and what you're tracking. The difference between precision or just knowing something is in the room is quite different, isn't it?

63
00:32:52.000 --> 00:33:14.000
<v Speaker C>Yeah, exactly. And it's interesting, you know, if you are actually calculating location, then it's quite often things are put against a wall. And so even calculating that something's in a room, if you're trying to have a continuous coordinate system can be quite hard because a foot error can place it on the wrong side of the wall in a different room, which can be very important in lots of applications.

64
00:33:14.000 --> 00:33:15.000
<v Speaker B>Indeed. Yes, yes.

65
00:33:15.000 --> 00:33:22.000
<v Speaker C>Although if you're already trying to figure out which room it's in, then there are different solutions for that and that's more solvable.

66
00:33:22.000 --> 00:33:35.000
<v Speaker B>Yes, but some of those have come back round again even in the last 10 or 15 years that where is the machine in the hospital thing has come back again and again it is a difficult problem to really solve, I think.

67
00:33:35.000 --> 00:34:22.000
<v Speaker C>Yeah, yeah, yeah. So anyway, it started to get into that, but then in 2005 I ended up being approached by Intergraph, who were really the second biggest geospatial company at the time after ESRI. They were 600 million in revenue and three and a half thousand employees or something. And they approached me to go and be their cto and at first I said no. We had ongoing discussions and after thinking about it a bit more, I kind of felt at that stage with Ubisense, it was quite dependent on the success, success of the hardware components, which really my thing. And you know, it was a big opportunity to turn down, to go and do that. So I ended up going to be CTO at Intergraph for a couple of years.

68
00:34:22.000 --> 00:34:24.000
<v Speaker B>Was that in the U.S. or.

69
00:34:24.000 --> 00:35:46.000
<v Speaker C>Yeah, that was in the U.S. they're based in Alabama and you know, I was already in the US at that time and actually sort of stayed in Denver and worked partly Remotely and, you know, but went down to Huntsville, Alabama fairly frequently and I had a good experience there, a lot of good people there. One cool experience was I got to talk at their user conference just two or three months after joining. There were 5,000 people there and General Colin Powell was the keynote speaker and I was on right before him. So that was a sort of bigger league or the biggest league conference I've talked in a bit nerve wracking, but it was a good experience and, you know, I learned a lot and it was interesting. They had good, solid products. But overall I kind of felt that I was a bit too removed from the technology. You know, I like inventing new things and being quite deep in the technology. And after sort of 18 months or so there was a point where Intergraph ended up being sold to or that they were publicly traded and they ended up being sold to private equity investors. And I decided that was a time to sort of move on. I was on the road almost continuously as well, so. So yeah, it was a good experience, but a little less technically involved and in most of my other roles.

70
00:35:46.000 --> 00:35:57.000
<v Speaker B>So how many, how many people were you leading in that, in that role? Was there a big technical department or was it more liaising with the, with this business units and the sales and all that?

71
00:35:57.000 --> 00:36:57.000
<v Speaker C>Yeah, so it's interesting actually, it was more liaising. I had, you know, the cto. You know, I've had three CTO roles and there are different, you know, quite different how people define that. You know, some companies have the CTO being over the whole development organization, they all report up to the cto. I was always kind of. And to answer the question, most of the time at Intergraph, I didn't have anyone reporting to me. I was more a sort of connection point between the different departments and I kind of would outwardly present the vision, work with the different product management teams and the development teams and support the sales teams. And that was similar to what my role was at Small World. Towards the end, they did change things. So the product management team reported into me. But most of the time it was more liaising between different groups in the company and that kind of thing.

72
00:36:57.000 --> 00:37:00.000
<v Speaker B>Yes. So where to after Intergraph then?

73
00:37:00.000 --> 00:37:33.000
<v Speaker C>Yeah, so after Intergraph I did my own thing for a while. I started a company called Spatial Networking. You know, one thing that I kind of was in the back of my mind was, you know, I'd. At that point I'd been CTO of two of the top 200 software companies in the world, but they were both companies that if you talk to somebody in the pub, they'd have no idea who they were. Yearning to do something more consumer focused, where it would be cool to say, well, I built this thing and it's something they've heard of,

74
00:37:33.000 --> 00:37:35.000
<v Speaker B>something your mother could understand, that kind of thing.

75
00:37:35.000 --> 00:38:20.000
<v Speaker C>Yeah. I developed an application called where youe Gonna Be, and I could have probably thought of a snappier name, but it was hard to get a dot com thing, but it was kind of interesting. I still haven't seen anybody do what we did, but it was about future location of your friends in particular. So by that time there were quite a few apps showing where are your friends right now? But I always felt if you randomly found out that somebody's in the same town right now, then it may well be too late. You've got commitments, but if you traveling quite a bit, you know that next week, you know, I'm traveling to Boston and Jeremy is coincidentally going to be there from Nottingham. If you know ahead of time, you can coordinate and meet up.

76
00:38:20.000 --> 00:38:21.000
<v Speaker B>Yeah, the future chicken.

77
00:38:21.000 --> 00:39:54.000
<v Speaker C>Yeah, so. So we did quite a lot of interesting stuff there, how to model that future space, time thing, how to get information where people were. And I think we built some cool stuff. But long story short, we didn't get enough traction with it. And I did other miscellaneous things in that time as well. You know, that was also where I started to get involved with Open source. So when I was looking at the technologies to use, I was quite interested in PostGIS in particular. You know, I'd always had a database background and come across that. And I went to the PHOSP 4G conference in 2007 in Victoria, which was chaired by Paul Ramsey, who is Mr. Postgis. And he was very kind with his time and spent some time chatting with me and answering questions. And after that I really got quite a bit more involved in open Source and used PostGIS and leaflet were probably the main things that we used in the development of the Wagon and B application. So yeah, that was really where I got engaged with Open Source and also OpenStreetMap to some degree. Somehow I got randomly connected with Steve coast and I'm pretty sure the first US OpenStreetMap mapping party was in my loft in Denver in 2008. Steve showed up. That was the first time I met him. And we rounded up a few people. And then I was recently involved with OpenStreetMap as well in those early years, talking at a few conferences and things like that.

78
00:39:54.000 --> 00:39:56.000
<v Speaker B>Exciting times, weren't they?

79
00:39:56.000 --> 00:40:02.000
<v Speaker C>The dawn of all of this activity yeah, yeah, exactly.

80
00:40:02.000 --> 00:40:16.000
<v Speaker B>And how did it end up that you kind of chaired the phosphorgy in Denver in 2011 then? Was that. Did you volunteer or you kind of pushed into it?

81
00:40:16.000 --> 00:41:27.000
<v Speaker C>Yeah, sort of indirectly. So a friend of mine, Erica Wolf, planned to chair the conference and they. And I sort of volunteered to assist to be part of the group, but then they were sidetracked through some other life circumstances where they were not able to be the chair anymore and so I sort of landed in the role. It was a lot of work, but it ended up being very satisfying as well when we ended up with 900 people from 42 countries, I think, which was the biggest FOS 4G up to that point. And I think people felt it was a good conference. So yeah, it was a fun experience and it was good just to get another way of getting more engaged with open source community. You know, I haven't really done direct development on open source products, but, you know, that was my biggest contribution directly, I guess, was organizing then.

82
00:41:27.000 --> 00:41:35.000
<v Speaker B>And I know we looked to it as something of a model when we were developing the 2013 event in Nottingham as well.

83
00:41:35.000 --> 00:41:52.000
<v Speaker C>Yeah, yeah. And I remember chatting with Stephen quite a bit and others in the lead up to that. So yeah, it was the phosphor G conferences were. I haven't been to one for a little while now, but, you know, I always really enjoyed those. A lot of energy and a lot of interesting things going on.

84
00:41:52.000 --> 00:42:15.000
<v Speaker B>So after this period of kind of exploring a variety of things, including kind of your own company there, what came next and what prompted that? Was it just that the spatial networking wasn't taking off as you. As you hoped or. Or was there an opportunity that came up?

85
00:42:15.000 --> 00:44:02.000
<v Speaker C>Yeah, so a bit of both, you know, Spatial Net. So I decided to go back to ubisense, which is what, you know, we were talking about the indoor location tracking. And that was a bit of a combination of, you know, we hadn't got the traction with spatial networking that I'd hoped. And then I felt there was an interesting opportunity and ubisense was still quite active in the utility GIS space, but as a services company, they acquired a couple of companies that were doing services on top of Small World. So they still had a good presence in that utility GIS market, utility and telecom. And I really felt there was an opportunity for a new product there. I'd been very impressed with Google Maps, which came out in 2005 and then had evolved this ecosystem of being able to do customizations on top of it, but I still feel that in many ways that's possibly the biggest single change in the industry in the time that I've been in. It was the arrival of Google Maps. It just to me changed it to where everybody knew about digital maps and could use them. It was so easy to use and so much faster than the traditional GIS products. And what I'd heard in passing comments from quite a lot of enterprise people is why isn't our expensive enterprise GIS that we spent millions on as fast and as simple as Google Maps? And so that seemed like an opportunity to say, okay, well let's see if we can do something that is as fast and as simple as Google Maps. So that was kind of what I wanted to tackle when I went back to Ubisense.

86
00:44:02.000 --> 00:44:38.000
<v Speaker B>Yes, Google Maps really changed a lot of expectations, didn't it? It pushed things more towards web presence and that expectation of slippy maps and live updating rather than waiting 30 seconds for your map. And then simply that UI and interactivity difference, all of that together made huge differences I think, and somewhat then made a two tier structure of the kind of people who followed that and then the current incumbents. I think that's more or less where that kind of Neo Paleo divide of the time came from.

87
00:44:38.000 --> 00:46:10.000
<v Speaker C>Yeah, I was just thinking that was that sort of time in those discussions where, you know, there were a lot of traditional GIS people or Paleo Paleo people, whatever, saying that this Google Maps stuff is not real gis, it's not what we do. And there was the term neogeography that was coined. I was never a huge fan of, but that seemed to be the term that stuck. And I was certainly in that camp saying, hey, you can do a lot of really useful things with that. And so we went back to Ubisense and it was kind of interesting because we built the first cut of the product based on Google technology. So we used Google Maps and the Google Maps API and then we actually for the back end used Google App Engine, which I think is still around in some form. And it was really kind of an early serverless system, but people didn't use the term serverless then. I was looking it up and it seems like that term got more widespread use in 2015, probably with Amazon Web Services and Lambda. But the idea of serverless for anyone who doesn't know is that you can just run code in the cloud and you don't have to worry about. It's not that there is no server, but it's that you don't know where it's Running it's just platform as a service. You upload some Python code to Google and it runs and it could be running on multiple servers or whatever. And so it was an early example of that kind of. Of thing.

88
00:46:10.000 --> 00:46:21.000
<v Speaker B>Yes. Not having to worry about having a piece of tin in the corner and managing all that stack of stuff that just connect you to the web. That was a big change, wasn't it?

89
00:46:21.000 --> 00:47:15.000
<v Speaker C>So I felt that was the future. And we developed this system that was basically Google Maps, but with utility data overlaid on it. So we'd sync the data from primarily Small World initially, because we had quite a lot of Small World customers. And later we added sync from ESRI and others. But so you have something that basically looked like Google Maps. It was as fast and as simple to use, but you could click and see information about your electric utilities. And over time we added more functionality, but really tried to keep it simple. But the interesting thing was people loved the functionality. But then there was very strong pushback about running in the cloud at that point, especially from utility and telecom customers who were a bit conservative on utilities especially. So they said, you know, we love the application, but we want to be able to run it in house. So we had to rethink about how to architect it.

90
00:47:15.000 --> 00:47:25.000
<v Speaker B>Do you think, looking back, do you think their concerns were good concerns or do you think they were just not up with the trend and just being too conservative?

91
00:47:25.000 --> 00:48:00.000
<v Speaker C>I think on the whole they were just being too conservative. You know, I mean that these days utilities are much more accepting of the cloud. And so I think it was largely conservatism and, you know, to a small degree, some valid concerns, that there was some sort of real time utility systems that was hard to move to the cloud because of specific security concerns. But I thought for this kind of thing that you could address, arguably address security better in the cloud than on their internal systems.

92
00:48:00.000 --> 00:48:03.000
<v Speaker B>But nonetheless it prompted an architecture change for you.

93
00:48:03.000 --> 00:49:22.000
<v Speaker C>Yeah, so basically we switched over to postgis as the back end at that point, which we could run in house, of course. And it initially used Google Maps on the front end, but then we increasingly had requests to be able to run offline, which involved doing sync from postgis. But Google Terms and Conditions specifically said that you couldn't run offline with Google Data or with the Google JavaScript code, even API. So we ended up switching the front end to leaflet, and Google was still typically used as the base map in most of our implementations and Google Search, but it was all wrapped in Leaflet on the front end. And then Talking to the customer and data in. In PostGIS on the back end. And then we added GeoServer into the mix for something. So we built this really very sophisticated enterprise app using a foundation of open source components. And at that point, in many ways I felt that was the main platform alternative to esri. That small world was getting on, Intergraph was getting on. So I think we were some of the leaders in building those kind of enterprise apps based on top of open source components.

94
00:49:22.000 --> 00:49:33.000
<v Speaker B>So did UB Sense do well? What prompted you to leave or change from ubsense in the end?

95
00:49:33.000 --> 00:52:48.000
<v Speaker C>Yeah, so Ubisense, well, ended up, you know, that product that we built there ended up doing well and Ubisense kind of changed or had a. The business kind of got split up in the. What's the time frame now? It's the sort of around 2017 or 18 or something maybe. So there was still the real time indoor location tracking part of the business and then there was more, the utility GIS part of the business, which had transitioned from services indoors into mainly product based on this new product that we built. So they were focused on quite different markets and the board decided to sell off the location tracking part of it. And it kept the name ubisense. And then IQgo was a new company formed with the geospatial side. And, you know, we had pretty good success. We got some large customers. One thing I was particularly proud of was that we had a top 10 telecom company in the US and every time they rolled out a new enterprise product, they did a customer satisfaction survey of their end users. And they'd been doing this for 20 years. And so, you know, I know Microsoft Outlook and Word and Internet Explorer, you name it, they all had these satisfaction surveys. And my world, which was the name of product, actually got the highest score ever in any of the products they'd rolled out in those 20 years. So that was, you know, very gratifying and a big vindication of what we'd done. But yeah, so I ultimately ended up leaving in I GUESS it was 2021. And I think, I know I just had a few disagreements on my role and things that I wanted to be doing and certain directions of the company. I had been very keen for a while on this concept of using computer vision to capture data, to keep the data updated, capture data in the field as it was installed and maybe come back to that momentarily. But anyway, I felt this was a revolutionary change for this part of the industry to be able to automatically update the data database and, you know, some others in the company management were less convinced about pursuing that or putting effort into it. Anyway, long story short, in some ways I got a bit pushed out and felt a little bit disappointed about that. But I ended up then approaching some other people about this idea. And so I went to work for. For SSP Innovations, where I had several friends working and they were sort of leading ESRI business partner, and they were intrigued by this idea and agreed to sort of fund some development on it. So that was sort of what triggered me to leave. And so, after many years of, in a lot of ways, leading the opposition to ESRI in that utility telecom space, I ended up joining an ESRI business partner. And I was told ESRI were quite suspicious. Suspicious that they had to confer with Jack before recruiting the case.

96
00:52:48.000 --> 00:52:52.000
<v Speaker B>I was just. In case we were there to spy.

97
00:52:52.000 --> 00:52:54.000
<v Speaker A>Yeah.

98
00:52:54.000 --> 00:53:10.000
<v Speaker B>And how did that go? So in that iteration, you were actually working on that AI data capture, was that successful in what you were trying to actually do with it?

99
00:53:10.000 --> 00:54:07.000
<v Speaker C>Well, so, yes and no. Long story again, I'm conscious we've got a long time here. I think we developed some things that I feel were probably the most innovative thing that I've worked on in my whole career. As we talked about, I've worked on quite a lot of interesting things. And every time we demonstrated it to people, they would be, wow, this is amazing. This is the future. This is incredible. And that he'd go back and carry on doing what they were doing before. And I think one of the challenges was that the ESRI customer base, which is the main customer base for ssp, are currently still very focused on migrating to the utility network, which is a sort of major upgrade of the ESRI product stack for utilities. And the older products are running out support. So I think to some degree that's just their top priority, above all else,

100
00:54:07.000 --> 00:54:10.000
<v Speaker B>being there a bit too early, in your case.

101
00:54:10.000 --> 00:56:45.000
<v Speaker C>Yeah. And then, you know, to some degree, I think it maybe didn't have as much sales focus as it might have had if it was a separate company. You know, that the core business of SSP was these large migration projects and the sales guys, you know, would make money off selling those. And this was a small new product and relatively inexpensive, so it could perhaps have had more sales focus. So it was a bit of a combination of the market being somewhat distracted and, you know, maybe we could have had more sales focus on it. So. But, you know, long story short, there, at some point, you know, SSP had been very supportive, but we reached a point where they decided, well, we haven't got more traction in the market, yet we're going to to not continue to invest in it. Which, you know, timing wise was interesting given that, you know, we started this prior to the emergence of ChatGPT and all these LLMs. You know, we were trying our own models and doing quite specific computer vision recognition and just towards the end of it, you know, the LLMs were coming out and we were looking at how you could use the LLMs and, and in some ways what we'd already built was better and in some ways the LLMs had advantages. But regardless, I still feel convinced that this is absolutely going to be the future. Where and very briefly, the whole idea is, especially in utility networks or telecom networks, you have people installing equipment in the field, you need to capture quite detailed information about how that's connected into the network and feed it back to the GIS to update system of record. And in general, doing it more manually is both a bit too complex and too time consuming for those guys who are focused just on installing things in the field. So usually someone else comes along after the fact and records it and that can be weeks or months afterwards. So there's this notion of an as built upgrade backlog, that the network data is kind of inherently out of date, whereas if you can just kind of scan it with your phone and capture it instantly, then that's to me a really game changing thing where you can keep the network immediately up to date rather than being multiple months out of date. And that has knock on effects on a lot of other business aspects of the business. And also as a side comment means there's a lot less need for GIS to do manual updates because in those businesses there's a lot of time spent manually updating the GIS network.

102
00:56:45.000 --> 00:56:56.000
<v Speaker B>And yes, if you ever wanted to move towards a digital twin representation of your network which is live state, actually understanding the complete current structure is rather a necessary thing it would seem to me.

103
00:56:56.000 --> 00:58:04.000
<v Speaker C>Absolutely. Yeah. It's a key thing for the concept of a digital twin and it particularly applies, as I was thinking about these things, there's a notion of updates that happen in the world that may not be known, about things caused by weather or whatever it might be, versus things that happen with intent, like an electric utility installing something new where you know that's happened and you just want a really quick and easy way to scan it and push it back in real time. So, you know, I definitely feel that that's the future for those industries, but it might have been another of these things where we're Just a little bit too early and I think maybe if we'd done it at IQGeo, if I'd been able to make it happen there, it might have got more traction. IQgo had a larger mobile customer base than Esri in general and SSP in particular and it would have been a more logical plugin to people who were using smart devices in the field already. So yeah, it was a bit disappointing that say we developed some really cool stuff and I slightly wonder whether to revive it in some form or not, but haven't.

104
00:58:04.000 --> 00:58:11.000
<v Speaker B>Yes, tell us where you are now then Peter, or what's happened in between and where are you now?

105
00:58:11.000 --> 00:59:25.000
<v Speaker C>Yeah, so I might be retired, I'm not quite sure. I haven't been working, doing paid work for the last year except very small things here and there. Been doing quite a lot of wildlife photography, which is a hobby of mine, but most recently I have been doing quite a bit with AI based development, you know, development background and I'm just blown away by what you can do with it. I've got into quite a bit of detail and have actually been developing an application for wildlife photographers that. It's quite a major data management problem. You know, you have hundreds of thousands of images and be able to identify species and figure out how good an image is. Has some kind of relationship to some of the computer vision stuff we were doing on SSP. P Vision, the previous product, but using LLMs but built something that would have taken quite a significant number of person years in a couple of months of effort with using Claude code. And so I am just really impressed with what you can do with that stuff. And that's going to have a very dramatic effect on all industries, including the geospatial industry.

106
00:59:25.000 --> 00:59:46.000
<v Speaker B>Yes, very much so, especially for people. And yes, I'm not sure whether it's more impactful for people who are graduating and coming into an industry about what they're going to be doing over their careers or people who are midway through their careers where suddenly all of this stuff is being disrupted in this way. I suppose time will tell on some of that.

107
00:59:46.000 --> 01:00:51.000
<v Speaker C>Yeah, I mean I find with the state of it right now that actually my kind of skill set where I am fairly technically detailed but I haven't done much hands on programming in recent years though, you know, I certainly have in the past, but I'm quite involved in product design and technical architecture and that kind of thing. It suits that kind of skill set quite well where you can sometimes just give a pretty high level instruction to Claude and sometimes dig in a bit more and say, you know, here's some technical suggestions or go in this direction. But you know that it's like working. So it's like working with an incredibly good team of developers. But if you have some of that technical knowledge and design skills and that kind of thing, I think that really helps. I mean, there's certain things you can do without it, but especially for building more complex systems with Claude, then it's helpful to have some of those skills. So I think learning how to work well with AI, I guess, is a good thing for all of us these days, but certainly for new people looking at what to do, that should be in the mix somewhere.

108
01:00:51.000 --> 01:01:17.000
<v Speaker B>Yes, indeed. So maybe that takes us into just rounding things off. What do you think of the geospatial landscape now? Part of the purpose of this podcast strand is to see some of these streams of activity and innovation coming together. So, for example, how special is geospatial, do you, do you think now from your perspective?

109
01:01:17.000 --> 01:03:13.000
<v Speaker C>Yeah, that's a perennial question over the course of my time in the industry, as I say, I think to some degree the answer is there. It's always, it's not 100% one or the other, of course. I mean, there are elements of it that clearly are special, but there's also a lot of elements of it that are in common with other types of data. And I've always lent a bit more to that side of the fence, I guess, going back to the old store everything a standard relational database. But then Smallwell developed a database that was very good at a certain thing, that was a bit more specialized and it made sense at the time. But yeah, in general, I think for most applications, the sort of broader geospatial universe of what you can do with open source products and indeed with Claude or whatever. I just used Claude last night for a trip I'm going on and I just tell it, read this itinerary, create a geojson file and a leaflet map for me, and it does it all. And so you don't need any real spatial geospatial knowledge for lots of things now. And you know, there were always people who would object to saying, well, what about thinking about accuracy and projection issues and that kind of thing? And to some degree, I think if you look at other types of data, like I used to say, you know, we don't have conferences on numerical information systems because there's an application that uses numbers, you know, and using numbers in an application doesn't make you a mathematician. And so while there are Quite specialized things in geospatial. There's a very large set of applications where you can do interesting things with geospatial data without being a geospatial person, just being a domain expert. And even more so with AI now.

110
01:03:13.000 --> 01:03:21.000
<v Speaker B>Yes, yes. Like a lot of things in life, it comes down to a bit of a spectrum rather than being a binary answer some of these things, doesn't it?

111
01:03:21.000 --> 01:03:24.000
<v Speaker C>Yeah, yeah, for sure.

112
01:03:24.000 --> 01:03:38.000
<v Speaker B>Okay, thanks, Peter. I think we've had an interesting discussion here. Thanks. And if people want to contact you afterwards, what's the best way for that?

113
01:03:38.000 --> 01:03:56.000
<v Speaker C>Probably LinkedIn is the best bet. I'm on there and you can contact me through there or Instagram and you can see some of my wildlife photos that I think that's actually how you connecting with me for this one. So those are both good destinations.

114
01:03:56.000 --> 01:04:26.000
<v Speaker B>Yes, I'm aware. Yeah, we've run fairly long here, but there are so many things I'd like to have brought into this. But I think you've had a fascinating career here of being at so many cutting edges. And I think part of the story seems to have been sometimes people appreciated that cutting, cutting edge. And sometimes it was maybe a bit in advance of either what your company mates or the market was prepared for.

115
01:04:26.000 --> 01:04:50.000
<v Speaker C>So, yes, congratulations, I'd say, I think another quote I remember somewhere, Mario Andretti supposedly said, if you feel everything's under control, you're not going fast enough. And I don't know if that quite directly corresponds. If you don't do a few things where it ends up that you were ahead of the market, then maybe you weren't pushing hard enough or maybe just didn't sell it well enough.

116
01:04:50.000 --> 01:04:57.000
<v Speaker B>I don't know. Yes. Okay, thanks, Peter. We'll close there. Thanks for joining the podcast.

117
01:04:57.000 --> 01:05:02.000
<v Speaker C>Great. Well, thanks a lot for having me, Jeremy. And sorry we weren't a bit long, but it was a good chat.

118
01:05:02.000 --> 01:05:07.000
<v Speaker B>Worth it. Thank you.

119
01:05:07.000 --> 01:05:41.000
<v Speaker A>Thanks for joining us today on the GeoMob podcast. We hope you enjoyed the discussion. You can get the show notes over on the website, which is@thegeomob.com and while you're there, you can sign up for our monthly newsletter and find links to our accounts on Mastodon and LinkedIn so you can stay up to date on all our upcoming events. We hope to see you at an event soon.
