Jack Dangermond Interview 3 of 3: The Geoweb

Jack Dangermond is the founder and CEO of ESRI. ESRI’s software is used by every level of government around the world. You can see ESRI’s influence in online mapping tools from Microsoft, Google, Yahoo! and FortiusOne. I had the opportunity to interview him over the phone on April 20, 2009. In this portion of the interview we discuss the history of GIS and online mapping.

Jack will be speaking at Where 2.0 on May 20th in San Jose. You can use whr09rdr for 20% off at registration.

Brady Forrest: What do you think will be coming in the next couple of years that will make the geoweb a reality?

Jack Dangermond: I think the geoweb is a reality. It’s just a matter of how you define it I guess. There’s just millions of pieces of content now available on the web for either FTP download and use in GIS or other analytic tools or as services that people are accessing by way of open systems. So I’d say maybe more analytic services that are available on the web or developers who take the foundations of technology as they exist today and build applications to them. But there’s got to be motivation. There has to be a commercial incentive for people like young developers to build applications on the web with web services. And maybe there has to be a marketplace created. I’m speculating here. I don’t have a clear answer for you on that.

Brady Forrest: You guys did have a marketplace for a little while, did you not?

Jack Dangermond: We do. We started with something called ArcWeb which was a little bit successful. It focused primarily on content and then developed an environment for people to build apps out of. But considering the cost of that compared to the kind of income that was coming in from it, it was not sustainable. Google and Microsoft’s websites in some way displaced that because much of the content that was there was free and it was much bigger. From what I’ve been told, somewhere between $75 and $100 million a year is spent on that content basemap to maintain it and serve it. That’s not sustainable if we’re talking about small developers. So we sort of shut ArcWeb down effectively. I mean we still have a few users that are using it, but it’s not something that sustains itself. If you have search and advertising revenue coming in, then you can afford to — as Google and Microsoft do, you can then afford to subsidize a lot of that content. But you can’t do it as an independent. So our second stage is –

Brady Forrest: I find that really interesting that you think the consumer GIS market is actually richer in some ways — on the web is richer than say the enterprise.

Jack Dangermond: Our second stage in the development of that is something called ArcGIS Online which is part of our actual software subscription, which has arguably the same kind of content and rich content that the consumer sites has as basemaps. And will I think iterate into a foundation for sharing a lot of the higher-end data content.

Brady Forrest: So when it comes to creating the open geoweb, what place do you think open formats have versus proprietary formats?

Jack Dangermond: Why don’t you further elaborate on what you’re getting at?

Brady Forrest: So KML has exploded in usage since it was not owned by Google anymore and became an open format and a lot more companies have adopted it. Sometimes GIS data gets locked up in different formats. It’s not readable by multiple platforms.

Jack Dangermond: So that’s a statement; isn’t it? It’s not a question.

Brady Forrest: I guess so. Yeah.

Jack Dangermond: Yeah.

Brady Forrest: What role do you think that has in creating the geoweb?

Jack Dangermond: I think KML is a very important standard. The other OGC standards are very important standards. As a vendor, we support all the standards and attempt to be open and standards-based with respect to our interfaces on the web.

Brady Forrest: Okay. And then just one final question. What do you picture is the future of interaction with spatial data beyond maps?

Jack Dangermond: Maps are a powerful to communicate spatial data and tell stories. So for me, maps are a kind of communication medium for storytelling. People want to tell stories; they use a map to be able to do it. And so for me, a lot of what I’ve seen in the consumer mapping world has been storytelling. I want to tell a story on conservation. I want to tell a story on where something is happen, where something’s cool, where there’s a great site for this or that. And I publish that and make it available for other people to see it. So beyond that, people are interested in analytics. They’re interested in saying, “Where should I live?” And do it beyond simply visualization. They want to be able to use a model like when I mentioned before Dr. Maidment good work. If I could overlay a lot of different layers in a model format and then manipulate it and say, “Here’s a very cool place to live,” that’s spatial analysis. That’s a model.

So I think even 45 years ago when we were first inventing computer mapping at Harvard, our first attempts were to make beautiful maps. They weren’t very beautiful; they were pretty crude. But very quickly thereafter, we moved from mapping into analytics, spatial decision-making to support land use planning and business planning of various sorts. So I think if you’re asking me where will the web and the geospatial world move, it will be from simple map displays to analytics on the web. It’ll move from simply a basemap with overlays on it to the notion of multiple distributed services that do spatial analysis, but combine various kinds of spatial data together and serve it back out as visualization communications.

Brady Forrest: Okay. Well, thank you very much, Jack.

Jack Dangermond: You’re welcome. Thank you, Brady.

This was the third portion of a three-part interview with Jack. The first was on Web Mapping and the second part was on Sharing Government GIS Data.

tags: ,
  • Archie Belaney

    So when it comes to creating the open geoweb, what place do you think open formats have versus proprietary formats?

    So that’s a statement; isn’t it? It’s not a question.

    Huh?

    Jack, you jumped on Brady for the audacity of bringing up a key point about ESRI’s presence on the Geoweb. And you never answered the real question (yes, it was a question, despite your assertion) that asked what is your take on open formats vs proprietary.

    Of course ESRI formats are proprietary. And of course KML is probably one of the largest threats your company has ever faced.

    Just as Autodesk and Microsoft have an ever-changing sequence of file formats to control and manage the data generated by their install base, so too does ESRI morph its formats to control access to the ‘world’s geodatabase’ that consists of all that proprietary ESRI data locked up in all those silos your customers have built.

    You didn’t answer the question, and instead browbeat Mr. Forrest into a corner, precisely because the issue is one you CANNOT address without acknowledging the strength of Google’s success with KML.

    How about a real answer? Wouldn’t that be an open and transparent discussion?

  • GeoGeek

    there’s the world of data format…
    and there’s the world of web protocols and web “formats”, that’s what we’re talking about in the interview

    ESRI’s Web solution is one of the most open one when it comes to web interoperability (that’s precisely Jack’s answer):
    – OGC’s services protocoles : WMS, WCS and WFS (certified)
    – Google and OGC’s KML : ArcGIS Server is completly exposed in its capacities as KML “services”.
    – Google Map and Microsoft Virtual Earth : ArcGIS Server comes with an extension API to those platform APIs to better integrate the GIS data and tools in the market’s visualization tools. Examples are coming showing integration of Open Street Map, …

    Who cares in what “format” are stored data exposed and accessible through open web services ? What we should really care about in this space is versability, performance, scalability and reliability, that’s the geodatabase story and that’s another story.

  • Archie Belaney

    OGC – HA! HA! HA! – the (certified) protocols are out of date, and require an adapter set to iterate correctly (otherwise known as the ESRI Data Interoperability Extension) and the WCS and WFS iterations are hugely unstable. Letter of the law, not even close to the spirit.
    KML services – well, yes. But the ESRI render-engine is bog-slow. Incredibly frustrating in comparison to other (freeware) tools.

    Versability, performance, scalability, and reliability – that’s the geodatabase story, and it’s not all that. If you want these things, go with Oracle Spatial, or other ROBUST and REAL database solutions…not some three-layer bloated fflat-file ESRI disengenuously calls a “GeoDatabase”

    Yes, ESRI in every count meets the letter of the law…see, we’re really kewl neat open folks just trying to help save the world, one map at a time…and in virtually every iteration the functions are woefully slow, unstable, or require umpteen hours of ‘special tuning’ to get them to work.

    There are SO MANY better ways to leverage the cranky, obfuscated, and ubiquitous content that’s locked up inside the ESRIverse.

  • Alan

    I also liked this recent interview with Jack
    http://www.govtech.com/gt/articles/652976

    Lots of good, current info there

  • http://gisconsultancy.com/blog Rob

    The animosity between ESRI and neo-geography is palpable, and alive and well in the transcript of this telephone conversation. It’s worth remembering both him and Hanke made a promise on stage at Where 2.0 last year to improve accessibility to ‘geodata’, and Jack stuck to his word, services published with ArcGIS Server 9.3 provide a KML footprint by default; whereas the pledge from Google to open up their search index of KML files sadly never materialised :-(

  • http://www.galdosinc.com Ron Lake

    I am surprised about the confusion in this interview re the word “format”. This word should ONLY refer to encodings that specify file structures (like a Shape files). We worked hard 20+ years ago to separate programs from data – that was one of the key issues in databases. XML also does that, so XML is NOT a format.

    More importantly there is no equivalence of these different encodings. It is nonsense to talk converting Shape files to KML or KML to GML etc. These encodings have different purposes and different rationales to exist. One can style GML to KML for visualization. One can encode GML features in a set of shape files etc. The abuse of language only contributes to misunderstanding.

    I think the ability to style data from ArcGIS Server to KML is very cool – but KML is NOT a data encoding – does not support content models.

    There is SO MUCH more to the GeoWeb than revealed in this interview. I was somewhat shocked by what it did NOT say.

  • Archie Belaney

    Perhaps the central issue here is ‘sharing’ information between platforms.

    One view might hold that you haven’t shared something until a data instance has moved from one location to another. AKA – here; take my Shp file, load it into your app, then post it on your site.

    Another view is sharing occurs when I publish maps that are readable by you – and once you’ve read the information and incorporated the layers/links in your aggregated view, then I/we are sharing.

    Yes, yes. This is very simplistic, yes.

    But this is the GeoWeb – some see it as a medium to improve tried-and-true practices, and some see it as a more free, open, and dis-aggregated collection of openly publishing and consumable nodes.

    I’d say Mr. Dangermond’s statements place him FIRMLY in the former camp, bucking the spirit of the open-consumption crowd as ‘not useful’ and ‘only dots on a map’

    Some years ago, I think it was Mr. Dangermond who said there are doers, users, and viewers. And that viewers make up the majority of the geocommunity, by far. Well, the geo-web is the perfect medium to serve the viewers…

    ESRI will continue to be the tools by which the doers and some users will develop the core content, and these products will be published freely and openly on the geoweb. That this marginalizes ESRI into