Straight Talk from the Top; Part Two

Part One of this interview is here

The future of ArcView

The plan is that the next generation of ArcView, which will be called ArcView 8.1, will be released simultaneously with ArcInfo 8.1; and it will be built from the same common technology we've already shown in ArcInfo 8.01 and 8.02. It will essentially be ArcMap and ArcCatalog, minus the ability to edit coverages and geodatabases. Our goal is to do everything ArcView 3.2 does, in the new framework. There will be a few exceptions to that, because they just don't make sense. But our goal is to get as close to that as is feasible.

We're working on that. What we know is that we can definitely convert some things, and we definitely cannot convert others. Things like symbology, pathnames we can do. Things like Avenue code we can't yet do.

We're working on a 3.2a patch right now, which we hope to have on the web, freely downloadable, before the User Conference. And we definitely envisage a follow-on release beyond 3.3.

Yes, half a million seats.

We're going to try. We also want to offer some other price points in between the $1,000 and $10,000 points, so that there's not a huge binary divide between the ArcView user and the ArcInfo users. We've been playing around with the notion of an ArcInfo Editor seat. Maybe an ArcInfo Map Viewer seat as well. This would give us four or five offerings in a spectrum that people can buy.

Other technologies from ESRI

The next generation of ArcExplorer will be Java based, and it will be cross platform. It won't have very much more functionality, just a few things, and it will continue to satisfy the low end.

Right now Windows CE and other PDAs have such very strict and very specific requirements. This means that we decided it's better just to develop technology for that environment than to try to port generic technology. You've got the constraints of lower speed processors, very constrained screen real estate, and the need to work with a pen-based user interface, and a number of other things as well.

Interesting Background and History

As you know, ESRI is not very big on titles. I usually go by Director of Products. My role at ESRI is really two-fold: one is to work with Jack to help organize and manage ESRI, our vision and our story. The second role is to work more closely on products with Scott Morehouse and Clint Brown. You probably know that Scott is in charge of software engineering and basically manages the programmers at ESRI. Clint's role is more involved with release, QA and documentation for the products. My role is managing the product managers, to do more planning activities.

We work really as a team, the three of us. We all try to do a little bit of everything. Jack, myself, Chuck Kilpack to a less extent these days, Scott, Clint are the main driving force in terms of what happens with products at ESRI.

I originally met Jack in 1986 or 1987 when I was a young lecturer, a University Professor at the University of Leicester. A couple of years later he asked me to open the ESRI UK office. I did open that office in 1990 and ran that operation until 1996. First of January 1997 I started working in Redlands.

What ESRI UK does, like many ESRI offices, is that it is the physical sales office for the United Kingdom; it is also responsible for adapting and customizing the software so that it meets the requirements of the local market and local users; things like translators back and forth between National Transfer Format, the local standard we did. Some of the UK customers were ahead of customers in the United States and elsewhere in the world. So we ended up doing some extensions to the core ArcInfo platform to support their requirements.

In the early days we wrote the prototype for dynamic segmentation. It was actually written for Cornwall county council in 1988. And then the person who did that, Val McDuff, actually worked in Redlands and helped work with the development team to help implement the first dynamic segmentation system in the software.

A second major project we worked on was the precursor to regions. We developed an AML design which we implemented for what was then called overlapping polygons. We designed a superstructure in AML which allowed you to have multipart overlapping polygons inside a single coverage with a tabular system running in the software. That was installed in several sites. And then that was used as the design input to create the regions subclasses on top of the coverage model, which was all done in Redlands.

There are a few other things as well, but these are probably the most notable; we always had a very good working relationship with the core development team in Redlands.

I think there are a number of reasons for that, even though there are more people in Europe than there are in America. There are many subcultures and many language differences, so Europe is really not a single large market but rather a collection of smaller markets. The American economics of scale effect makes it much easier for companies to develop to a large size in America. We always used to think that in Britain, where the aim was to be about -- we could build a market of about one-tenth the American size. So that if our turnover and size is about one tenth the size of the equivalent American company, we were about on par.

I think also, to fully answer the question, there are some significant and very notable European GIS companies. including Smallworld, Laserscan, Siemens in Germany, Star and Geoconcept in France, and others as well. They're reasonably important at a national level, but at a global level, all of the major global players are American based.

Yes, but if you look at them, a lot of their success has come from having a strong presence in the United States.

Well, personally, I definitely look at what other companies are doing technically and what they're thinking in terms of the way they're organizing things. Clearly Microsoft is a major player. We try to track what Oracle are doing and what they're thinking and how they're organizing things. Sun as well would be another company. We've been tracking SAP a little bit in the past couple of years because we've had quite a close involvement in working with them. These are the main companies.

The only company that is at all comparable to us in terms of what they're doing is SAS. By coincidence, they're also an institute, the SAS Institute, as we are the Environmental Systems Research Institute. They're a private company, totally privately held. They're in the mainstream of IT; they're in statistics and statistical analysis and business intelligence and reporting. They are a similar sort of size to ESRI, give or take. Probably not more than $100 million difference either way.

Yes, I think it's scaled pretty well. The way ESRI is organized is we're a collection of small teams organized into a series of divisions. That's a fairly scalable model, because we can always add more teams. It is challenging at the management level to manage an increasingly advanced technology, and increasingly large teams, but so far we've been able to pull it off through the ArcInfo 8 thing, which is the biggest GIS project ever undertaken, inside ESRI or anywhere else.

Part One of this interview is here