00:01Now I have the big privilege to actually introduce to...the next speaker, this morning, Michael Goodchild.
00:07He definitely doesn't need further introduction that his name, his name has been very, for a long time, related to GIS.
00:14And for the last few years he has been one of the big advocates in actually trying to merge design and GIS.
00:21He has organized and hosted a number of events back in Santa Barbara where he's a professor.
00:27In sessions, brainstorming sessions, in expert sessions dealing with issues like How can we integrate design and GIS?
00:34And what are the concepts that sustain such integration?
00:38So without further delay, I would like just to welcome Michael Goodchild; please welcome with applause.
00:50Thanks, Juan Carlos, and thank you for the invitation to do this...
00:53...and thanks to Jack and Esri for being willing to host this, this meeting.
00:59It's great to be on the summit; the view is tremendous.
01:04I was thinking of this coming in this morning, looking at Mt. San Bernardino and Mt. San Gorgonio...
01:10...and thinking how wonderful it would be to be up there, to be able to see so far.
01:14And I think that's one of the wonderful things about a meeting like this, that we can see so far and see the vision of what might exist.
01:24But one of the things I have to remark is that I think the vision is enormous.
01:29It's as big as the horizon that one can see from the top of Mt. San Bernardino.
01:34And coming to grips with that is going to inevitably require us...
01:37...to come down off the mountain and to descend to a more practical level.
01:42And I want to do that to some extent in this talk by talking about GIS and where GIS has come from...
01:50...and where it currently is and how it might serve the needs of design better.
01:56So this is perhaps the beginning of a discussion that will go on for the next two and a half days...
02:02...on exactly what we do to move this agenda forward.
02:06So I think this takes us in a different direction from the previous two talks...
02:10...which I think provided a wonderful, motivating context for this, for this meeting.
02:15And I'd like now to look at the GIS issues quite specifically.
02:20So let me start with my interpretation of the GeoDesign vision...
02:25...and this is pretty much consistent with all the material that's on the Web site...
02:29...and with some of the discussions that have led up to this meeting.
02:33And essentially it has two parts, and I think it's important to emphasize the, the relationship between these two parts.
02:40One part is about input and editing and recording, and the word sketch is used very often.
02:47The idea that the user might be able to input a sketch of an idea, to have that input into the system, to do this collaboratively involving...
02:59...and there's talk of millions of people being involved in this process.
03:02To do it from different kinds of devices, from very sophisticated devices and very primitive ones...
03:08...to transform those sketches into features, to add them to a geodatabase...
03:13...and this is very much what ArcSketch has been building towards...
03:18And it's also something related to, for example, Google SketchUp.
03:22There's technologies of this nature out there.
03:26But then, the other half, and to me, this is very much the other half...
03:30...and that's the half which allows those sketches and ideas to be evaluated, analyzed...
03:37...to use prediction to see what the consequences would be, to modify them, to improve them.
03:42To do this according to well-defined procedures.
03:46And this, of course, is something that GIS is tremendously powerful at doing.
03:51We have in GIS an abundance of the kinds of tools needed for that process.
03:57So in some ways I see that this meeting is trying to bring these two topics together.
04:03Bringing together the idea of interaction and sketch and, and idea creation with evaluation...
04:10...based on the knowledge that's been accumulated in many disciplines.
04:15Now if this sounds familiar, of course, it is, and it's very much taking us back to the world of Ian McHarg...
04:22...and his school at the University of Pennsylvania...
04:25...because much of what McHarg was trying to do in that period was along these lines.
04:30It was using knowledge of meteorology, geology, hydrology, plant ecology, animal ecology, limnology, and computation...
04:38...and remote sensing to build the kinds of tools that would be needed to achieve that vision.
04:46And so one of the things we might do, and this, in fact, is, is a slide from a presentation that Jack and, and Carl Steinitz, and I...
04:53...made at the National Science Foundation in 2003.
04:57One of the things we might do is to try to move forward that McHarg vision...
05:02...and move it into the context of 2010 and see what, in fact, what has happened to that vision.
05:07So let me just briefly focus on, on what has happened to that concept from the 1960s, the design with nature concept...
05:15...because McHarg talked, of course, about layers, and at the time, his primary mode of operation was transparent layers...
05:23...superimposed on a simple light table.
05:25And it's easy, of course, to see how that has led, over the years, to the layer concept that underlies GIS.
05:33So in many ways, what we're doing with GIS today is an implementation of that 1960s idea which, for, of course...
05:40...a variety of obvious reasons McHarg wasn't able to, to implement to the kind of, of level that we can today.
05:48And here's a quote from a wonderful book, and I don't know if any of you have had a chance to, to read it.
05:53It's McHarg's autobiography. It was published by Wiley in 1996; it's called A Quest for Life.
05:59And he says, "For the first time, a department of landscape architecture could recruit a faculty of distinguished natural scientists...
06:04...sharing the ecological view and determined to integrate their perceptions into a holistic discipline...
06:09...applied to the solution of contemporary problems."
06:12I think that still stands as very much what underpins why we're here.
06:18That same thinking, I think, is very much here today.
06:22What he was talking about was integrating science into action.
06:25Integrating the knowledge that we have in a variety of disciplines into intervention and action in the community.
06:32This has frequently been emulated, but very often that intervention component has been weakened.
06:38And Jack will remember, in 2003, when we made this presentation at National Science Foundation...
06:45...and suggested the National Science Foundation might foster this kind of thinking.
06:50The first response we got was from one of the audience who raised his hand and said...
06:55..."That's the scariest thing I've ever heard in my life."
06:58The idea that you would take science and try to use it in a practical context...
07:03...was something quite alien to the basic scientists of...of the National Science Foundation.
07:08Moreover, I think the social context of this is missing.
07:12And it's something that today we would have to take much more seriously.
07:15And finally, computation and remote sensing, very primitive in the days of the 1960s, today, of course, are much more powerful.
07:23So if I were to try to move this forward, and this, again, is a 2003 slide, it would suggest this...
07:29...that computation and remote sensing are now an inevitable part of all of this.
07:35David Simonet and Waldo Tobler, in fact, were advisors to Ian McHarg back in the 60s.
07:39Bruce McDougal was hired; he was an author of a very early text in geographic information systems.
07:45And technology became a source of data, an engine for computation, a means of visualization.
07:50And it provided a framework that was formal and replicable.
07:54Something that could be defended in court...
07:56...something that could be shared between people because we shared an understanding of what it was trying to do.
08:03So 35 years later, this is 2003, has the science of intervention evolved?
08:09Has this evolved into something that we would recognize as the science of intervention?
08:14I think the answer probably is no.
08:17We have not achieved perhaps what, what McHarg hoped we might achieve.
08:21Has intervention become more scientific?
08:23Again, a debatable question, something we might want to debate here.
08:27Has the role of technology advanced?
08:30What are the components of that technology?
08:31Well, GIS is a very clear and recognizable component.
08:36And how should we update the McHarg model?
08:39So let me just spend a couple of minutes on that.
08:42The McHarg team of 2003; this is in contrast then to the McHarg team of 1965, 66; it would contain these days information scientists.
08:51We might call them geographic information scientists; they'd be concerned with information integration, information management...
08:57...semantic interoperability, visualization of scenarios, spatial decision support systems, public participation GIS.
09:04All of these things would have been alien terms in the 1960s, but today would be an inevitable part of that McHarg vision.
09:11Moreover, we would involve the social sciences, I think, and provide a much richer social context to all of this.
09:17So we would involve decision scientists, concerned with uncertainty and risk.
09:22We'd involve cognitive scientists, concerned with the design of human-computer interaction, treating IT as an enabling technology...
09:29...not imposing itself on the process.
09:32We'd involve social psychologists, who'd be concerned with the process of group consensus.
09:37So the science today underlying that McHarg model is much richer than it was in the 1960s.
09:44And we would intervene at, I think, a different scale.
09:47We would involve environmental economists.
09:49We'd involve political scientists in the process.
09:52So that's taking that McHarg concept from design with nature in the 1960s and moving it forward.
09:59Now let's focus on GIS, because meanwhile, GIS has been developing.
10:03GIS, over the past four decades, has become a technology for automated cartography, a technology for measurement...
10:11...a technology for management of assets, and for scientific discovery.
10:16But besides those, McHarg's vision is still one of the roots of GIS.
10:22The idea that GIS is a technology for design is there very much in parallel with GIS as a technology for, for example, management of assets.
10:32But at the same time, I'd suggest that the McHarg vision has somehow got lost along the way.
10:37We have become busy in GIS doing things, other things, with GIS, things other than design.
10:45So that today, I'd suggest that seeing GIS as primarily a design technology is somewhat unusual.
10:53And instead, I'd, when I teach about GIS, I teach about things like managing assets...
10:59...managing the assets of a utility company, for example.
11:01Very different from the design context of McHarg.
11:05So perhaps one way of seeing the business we're at here is in redressing that balance...
11:12...bringing GIS back into a more design-oriented technology.
11:17So coming back to my two parts, I see GIS as, I see this GeoDesign context, then, as having two related parts.
11:26The first part is sketch and record, user interaction, sketching ideas...
11:31...and the second part is evaluate, analyze, predict, model, improve.
11:36I struggled to find a nice, convenient acronym for the right-hand side.
11:41I couldn't find something that was pronounceable; I tried dropping some of the letters and substituting others.
11:47I'd suggest we make that one of our tasks for this, the next three days.
11:53I, I thought of referring to this as the yin and yang of GeoDesign, the left side, the right side, the reds and the blues.
12:01Somewhere there is an elegant way of expressing this...
12:06...that there are two interrelated parts that we must consider in our discussions here in the next couple of days.
12:11So taking that right-hand side, taking the yang, if you like, what do we know about EAPMI?
12:17What do we know about the yang of GeoDesign?
12:19What we know, I think, is that ArcGIS already has many, many tools that do many of these things.
12:27But these tools typically are in isolation, and they're not integrated with the sketch and record side of the yin and yang.
12:36We have the world of spatial decision support, and I want to credit Naicong Li and her group, who've done, I think...
12:44...a tremendous job in building the Redlands Institute SDSS portal...
12:48...which is a wonderful resource for spatial decision support.
12:52But this is still a little short, I think, of what we're here to discuss, which is a much more engaged process that involves the community at large.
13:03SDSS still remains, I think, a technology of the expert.
13:08So let me just cite a few examples, because these are the tools that already exist in ArcGIS for design.
13:15And perhaps what they will do is illustrate what a broad canvas we're actually here to discuss.
13:21The horizon is tremendously broad.
13:24So vehicle routing and scheduling, for example, we have numerous tools in ArcLogistics for designing, bus routes, delivery routes.
13:31We have numerous tools for optimizing travel on networks, minimizing fuel used, minimizing time, et cetera.
13:39So just a couple of quick examples. Here's a ArcGIS application for the problems faced by Schindler Elevator.
13:48This is designing their daily workload in downtown Los Angeles.
13:52Optimization to minimize the amount of time spent traveling between, between sites, the sort of thing...
13:59...the sort of design task that GIS can already do very well.
14:03Here's Sears, another client of Esri that uses the same kind of technology.
14:10There's one area where design already is there in ArcGIS.
14:15Location and allocation, finding the best locations for facilities that serve dispersed populations.
14:21Optimizing store sales, minimizing distances traveled, minimizing construction costs.
14:27All of them very much design but very much focused on infrastructure, very much focused on business.
14:33Very different from the kinds of examples we talked about in the first session today.
14:37So here, for example, this is actually a competitor; this is GE Smallworld, being used to optimize the location...
14:43...design the locations of Tesco stores in part of Britain.
14:47Here is some work I did 30 years ago on school districting in London, Ontario, again, using GIS to optimize the design.
14:56And we have abundant technology for locating linear facilities, pipelines, highways, railroads.
15:03Optimizing environmental impacts, construction costs, operating costs.
15:09Just a couple of examples; this is something I did recently...
15:11...in the context of Native American, pre-Colombian populations in Southern California.
15:17This is trafficability; this is the central point is roughly the location of the Santa Ana Airport, and what it illustrates, for example...
15:25...is how easy it would have been for Native Americans to have passed through the Santa Ana Gorge into the Inland Empire.
15:34We can do this kind of thing with existing tools.
15:37Here's a wonderful example of using GIS for optimum design.
15:42These are the wildlife bridges over the Trans Canada Highway in Banff National Park.
15:47Very successful as ways of allowing wildlife populations to cross the four-lane highway without accident.
15:55And we've done a lot of work using GIS for land-use modeling.
15:59So, for example, we have worked on predicting urban growth, predicting land-use transitions.
16:05We have models that users can use to control parameters, control in, in, initial conditions, but expert users, that is...
16:13...experts who understand the models and understand how they work.
16:16So here, for example, is the work of Keith Clark, my colleague at Santa Barbara...
16:20...forecasting the development of the Santa Barbara area under various growth scenarios.
16:26A simple model, it's actually a cellular [unintelligible] that is based on simple variables...
16:32...such as existing land use, elevation, slope, and access to transport.
16:39Here's a model from the, from British Columbia, from the lower Fraser Basin, the quest model.
16:45This is much more geared towards public participation because the public is able to modify the parameters...
16:51...through a simple interface and then see the consequences in terms of predicting growth in the, in the Fraser Basin.
16:58So these are things we can do.
17:00But to come back to my earlier point; these are not integrated with the other side of the yin and yang.
17:06They're not integrated with sketch and record.
17:08They don't allow the non-expert user to participate in the design process.
17:13What they do, however, is something that I think is very much a part of this discussion and something that, in many ways...
17:21...captures what I visualize as the future of this technology.
17:25What I would like is to be able to bring up a map such as this, to design by introducing sketched features...
17:34...and then to press a button and have that button evaluate or predict or improve on my design...
17:42...by bringing the strength of GIS and all the models that we have available...
17:46...and all the scientific knowledge that we have to assess, model, and analyze my suggested solution.
17:54That's something that I think is all too often missing.
17:56It's something that we can conceive of doing given the strength of GIS in the background.
18:02So that's why I think this yin and yang is a very appropriate way to frame this discussion and to think about it.
18:09Just a, a few other topics I should mention briefly.
18:12One is devices, because I think part of this vision is that these solutions would be interactive; this would be an interactive technology.
18:21It might use physical analogs, and I'm sure we will see over the next couple of days, physical analogs including that table.
18:30I know, for example, the work of Leo DeSilva's group, which has involved the concept of...
18:35...of moving clay in a virtual environment...
18:39...actually shifting clay around, shifting various kinds of physical analogs around in a virtual environment.
18:48We can think of this as high end and low end; we can think of the high end as being virtual cave environments or tables.
18:55We can think of the low end as being nothing more than cell phones and Wii terminals.
19:00So a huge variety of possible devices might be used to implement this kind of vision.
19:05So here's just one example; this is actually the work of Antonio Camara's group in the University of Lisbon...
19:11...and this involves a table, and virtual features are being moved on the table.
19:16To, to go back to my vision, what I would like to see in a table like this is a set of buttons that I can press to bring all of the tools...
19:25...and all of the power of GIS and scientific knowledge to bear on assessing, analyzing, and modeling my proposed solution.
19:34And then another element that I'd suggest is something we have to include in this discussion is the wisdom of the crowd.
19:40We have to think about this not just in then, in the sense of the top-down process involving experts...
19:46...but also a bottom-up process involving millions of potential stakeholders.
19:51And that's something that these days, we can achieve.
19:54We have the crowd sourcing technologies, we have Web 2.0, we have mobile phones.
20:01This is the world of neogeography, and if you haven't come across that term before, it's a wonderful term.
20:07It refers to the ability of the average citizen these days to do many of the things...
20:11...that geographers have traditionally regarded as their professional expertise.
20:16It makes me a paleogeographer, which is a wonderful thing to think about.
20:22I don't know if you've seen anything of the MIT, the, the winning...
20:27...the competition that DARPA ran a couple of weeks ago that MIT won...
20:32...which involved crowd sourcing to solve the problem of finding a series of red balloons located across the United States.
20:40And MIT's solution was simply to very rapidly recruit a network of people, a network observ...
20:46...of observers across the country and to use the crowd to solve the problem.
20:51It's a very elegant kind of solution.
20:53I'd suggest that's part of what we have to talk about here...
20:56...the possibility that today's technology can include a vast array of potential stakeholders.
21:03So to try to pull this together, let me, let me ask the question Where do we stand?
21:07And I'd suggest that we have currently some of the tools needed to achieve this vision.
21:14Some of them are integrated in GIS, but they're generally scattered.
21:19And generally they're not integrated with the other side of the equation, with sketch, with crowd sourcing.
21:25They're not integrated with the part, kinds of participation that we can now achieve through interactive devices.
21:32And what's more, the set is not complete; there is some major holes, major gaps in the set.
21:38So I think these are the kinds of questions we ought to be thinking about.
21:42We have some devices, but the interoperability between those devices is very limited.
21:48People who have worked with things like tables have typically developed their own software unique to that device...
21:54...and have not integrated across a variety of different devices.
21:57And we have very few studies of how users react to these kinds of technologies...
22:02...what kinds of design criteria they want to see implemented to make the interaction as, as easy as possible.
22:09So what needs to be done?
22:10Here's my suggested set of ideas for what we can talk about in more detail over the next couple of days.
22:17Number one, I think it's important that we try to map out all of the use cases.
22:22It's easy to become focused on some limited problems of design.
22:28And I think we need to keep our horizon very wide and try to think of all the different kinds of design problems that we will face...
22:37...particularly in the context of Tom Fischer's talk, which raised a host of design issues.
22:43And somehow I think we need to enumerate what those are.
22:47Number two, I think it would be impossible to approach this problem holistically.
22:53The set of possible problems is so large...
22:56...it will be very difficult to be wise enough to design something to respond to all of them.
23:01So I'd suggest the strategy we need to use is a strategy of rapid prototyping.
23:06We need to select a few problems and prototype what our vision means for those problems.
23:13We need also, I think, to integrate new kinds of user interaction; this means sketch, this means new kinds of devices...
23:19...and this means Web 2.0 kinds of concepts.
23:23We need to learn from those prototypes, and I'd suggest we need to learn from those prototypes...
23:27...particularly in the sense of the reactions of users.
23:31It's the ability of users to interact with these technologies...
23:35...which is going to determine ultimately whether they get adopted and whether they get used.
23:39So that has to be an important part of the agenda.
23:43And out of this, I'd suggest that we can hope that a comprehensive solution would emerge.
23:50I don't think a comprehensive solution should be designed top-down; I think it's too early to see the wood for the trees.
23:57So, thank you very much for your attention.
Spatial By Design: Understanding the Special Role of GIS
Michael Goodchild, Professor of Geography, University of California, Santa Barbara, presents “Spatial by Design: Understanding the Special Role of GIS” at the 2010 GeoDesign Summit.
- Recorded: Jan 5th, 2010
- Runtime: 23:07
- Views: 41353
- Published: Aug 25th, 2010
- Night Mode (Off)Automatically dim the web site while the video is playing. A few seconds after you start watching the video and stop moving your mouse, your screen will dim. You can auto save this option if you login.
- HTML5 Video (Off) Play videos using HTML5 Video instead of flash. A modern web browser is required to view videos using HTML5.