Simon’s Blog

There have been reports on the Net today about Sun co-founder Bill Joy’s revelations that Sun and Apple have more than once, through the years, come close to merging. In fact, Joy said, there were six close encounters. In the late 1990s, I was a strong advocate of such a move. Why? Well, it’s a long story. But the answer can be encapsulated in one word: Java.

In 1997, I joined a company (a Biotech that had just IPO’ed) that was committed to the Mac on the desktop, taking responsibilty for advanced aspects of IT in the business. Shortly after I arrived, we took the decision to invest millions in the development of novel, advanced computer systems aimed at transforming productivity and quality of core operations (drug discovery and development) in the business. My role was: to define where we needed to be in terms of IT; to come up with the strategy that would take us there; and to deliver on that strategy. Executing on the strategy involved us developing sophisticated, distributed computer systems that could be accessed by anyone in the company via powerful, easy-to-use software on the desktop.

Now, I’d been involved in the very earliest days of the web, including early web application development; which meant that I knew very well what both the benefits and the limitations of the browser were as a platform for application developement. While CIOs everywhere, mostly new to the web, were thinking about moving “everything into the browser”, my view was that, in situations where you had control of the desktop, there were other - more effective - ways forward. The point was: in the late 1990s, web applications were about ten years behind what could be achieved with a rich graphical user interface. The only advantage that the web app offered over the conventional app, then, was zero-admin automated deployment. And zero-admin automated deployment is simply a software problem; there’s nothing special about the web browser in that regard. And actually, the browser wasn’t, and isn’t today, particulaly smart when in comes to app deployment. For example, browser apps are downloaded each time the app is accessed - regardless of whether a new version is available.

I was confident that, through smart software - written by my super-smart team - we could have all the benefits of the web, without any of the limitations of the browser. The question was: what language should we use for our multi-million software development investment? I was concerned about locking our investment in to the Mac platform. The reason for my concern was that the day after I joined the company, colleagues were asking me why we couldn’t switch to Windows. However, I was in no mood to force a wholesale switch of platforms on the desktop across the company. While it would have been easy to make an argument for such a switch, I took the view that it would have been of little genuine help to the business, strategically; and making the switch would have consumed time and resources (both of which are in short supply in a rapidly expanding company). However, while I knew that we could hold off making a desktop OS switch for at least some time, it was also the case that a single M&A event could have have forced our hand, and require us to move to Windows on the desktop.

So, there was a dilemna - what if we sunk large amounts of money into developing software for the Mac, and then had to dump it due to a desktop platform change? Fortunately, to our rescue came the as-then relatively new Java platform (in addition to our other technology focus areas of CORBA, XML and Oracle). Even in the late 1990s, Java was a fabulous platform, and it was more than possible to develop apps that ran on the Mac, Windows, and Solaris (our development, and server platform) without modification. Java ran really great on Solaris, and really great on Windows. However, to be honest, it didn’t run that great on the Mac. A big part of the problem was the then-current version of Mac OS, Mac OS 9. The old-fashioned OS simply couldn’t cope with the modern, multi-threaded Java platform. Still, we had great success . Plenty of people didn’t, though: on what seemed like a daily basis, people from other companies were telling me that “Java didn’t work” or that “Java was slow”, or that “Never mind write, once run anywhere, it’s more like write once, debug everywhere.” What can I say? That wasn’t our experience. I guess there are lot of clueless developers out there.

It was obvious to me that the kind of success we had had with Java could be enjoyed by everyone. However, while the potential of Java was clear - that was, to become the dominant platform on both the server and the client - it was also clear that Microsoft wasn’t going to take the new pretender lying down. To make Java truly dominant on the client, my view was that Sun needed a serious desktop play (we’d had a free JavaStation from Sun - and, trust me, that wasn’t a serious desktop play!). So… what’s the next best thing to the number one desktop platform on the planet (Windows)? Answer - the number two desktop platform on the planet: the Apple Mac.

If Sun and Apple had merged in the late 1990s, they would have had a great shot at taking over the IT world - via Java. The merger rationale would have been to build the best server platform for Java (on Solaris); together with the best client platform for Java (on Mac OS). The combined organisation, I thought, could have been formidable - with a bit of invesment, the then new Mac OS X could easily have been the best desktop OS for Java.

Of course, as we all know, the merger didn’t happen: and the Mac never became a first-tier development platform for Java. At the time, I didn’t know Sun and Apple had been considering merging. I wonder what the world would have looked like today if they had gone ahead….