It is always an honour to interview people who have ‘served’ and worked on operating systems at the “golden” times of the operating systems, the ’80s and pre-Win9x days. Today we interview Adam de Boor, who was the CTO at GeoWorks, developers of the GEOS, in the begining of the last decade. Adam today works for OpenWave Systems. We discuss about GeoWorks, its past, its future, where it should have been.1. Please tell us about the great GEOS days. Do you go back to these days with a romantic/nostalgic eye, or was it business as usual? How important GEOS was for you?
Adam de Boor: The development of GEOS was a great time in my life. Just out of college, working 40 hour days (all-nighters a few times a month). We thought we were doing something great for the world. There were 16 million older model PCs out there that couldn’t run Windows 2.x (and certainly weren’t going to run Windows 3.0), and we were going to bring the ease of use of the Mac to all of them. I was one of the four architects on the team, responsible initially for development tools and some of the low-level device access. Gradually took on the file system, interprocess communication, virtual memory, networking, etc. Wrote a Macintosh specific user interface that got Claris interested in using our stuff (but that deal ended after a few months of
We had a really good team that worked together extremely well, and cared for and trusted each other.
I learned a heck of a lot about PC hardware (the OS was written entirely in 8088 assembly language, which I’d never used before coming to Geoworks), and operating system design. So yes, it was very important for me.
2. What went wrong with GEOS? Was Microsoft (again) one of the reasons for its dismissal from GeoWorks Corp as the desktop environment of choice?
Adam de Boor: It was a combination of factors, I think. Geoworks never gave marketing the respect it deserves, from the beginning through its demise. This led us to miss an important point: people who had old PCs at that time were the sort of people who were unlikely to buy something new, like Ensemble. They had the applications they were using, and it suited them fine.
In addition, we were missing two vital pieces when we launched: a spreadsheet, and an SDK. The spreadsheet was a very important business tool, and the SDK would have allowed us to actually create a developer community. Even without an SDK, there were several people in Germany who figured how to create an application without any of our tools, which I found absolutely amazing.
Finally, there was the major advantage that Windows 3.0 had when it launched 6 months before Ensemble 1.0: you could multitask the existing DOS applications using it. It might have been buggy. It might not have always worked. But it worked well enough that 3.0 was mostly used for task switching and solitaire, and that was enough.
3. What do you think about Breadbox’s Ensemble kit? Do you think it has a place in some niche market?
Adam de Boor: I still think the Ensemble applications are world class. Easier to use than many of the apps out there. When I was creating our wedding invitations, we were printing them on an orange vellum, and my wife insisted the text be in brown ink. Even today, you cannot get Microsoft Word to put your text in brown. But I could in GeoWrite.
With the various things Breadbox has added to Ensemble over the years, to interoperate with the Microsoft Office apps, have a good browser, etc., you’ve got a good application suite that still lets you interact with the rest of the world. I’ve not tried it recently, but I’d be surprised if the installation required more than 50Mb of disk space, and it still runs blazingly fast on 386 and 486 machines that folks still have running.
So there are a number of niches into which it can fit, largely having to do with education (which is usually underfunded), and programs in so-called developing nations, where refurbished older-model PC’s are starting to find their way.
4. Technically speaking, what was the strongest point of GEOS at the time?
Adam de Boor: GEOS had two major strengths, technically speaking:
1. it was highly, highly optimized
2. so much of what an application needed to do was already implemented as object classes in the system, with a nice model/view/controller mechanism to allow you to incorporate the classes into your apps.
We often pointed out that GeoWrite, which could do a number of things in 1990 that Word only got to in 2000, was only 150K, and about half of that was precompiled object instances. What we didn’t say was there was another 500K of code in the system that implemented most of the functionality. GeoWrite itself was primarily a document manager and UI resource.
5. Some say that the virtually impossible to crack boot disks of GEOS in the C64 was one of the reasons that GEOS never seriously took off (the 1541 drive was destroying the disks often, leaving the user with no way of booting the OS). What do you think about this?
Adam de Boor: I’m afraid this was before my time. But be aware that even as of 1998, Geoworks was still receiving royalties for the sale of Commodore GEOS product and SDKs. Commodore GEOS was, in fact, in the top 3 desktop operating systems for quite a while.
6. Do you think that there is space today in the operating systems market to create a desktop OS that can compete well with Microsoft Windows?
Adam de Boor: It’d be very difficult to do, due to the need to interoperate with many people with Windows desktops, and to deal with the large range of hardware that’s out there. I was the compatibility czar for Geoworks, and I know how hard that can be. Keeping up with printers alone absolutely requires printer manufacturers to develop the drivers, and they’re not likely to do that for
an OS without significant market share, which it’s tough to get without support for all the latest hardware. Catch-22.
7. If there is one thing that GeoWorks should have done to make sure that GEOS would continue its legacy and success, what that would be?
Adam de Boor: Making the source freely available would have been good. There are lots of interesting concepts in GEOS that I think people could learn from, and would have allowed interested parties to possibly create a groundswell of support for some of the niche markets.