Oracle paid $7.4 billion to buy Sun Microsystems, the formerly high-flying maker of server computers running Unix, back in 2009. Many wondered back then whether Google was in the running. They certainly had the cash, though the big question was why Google would want a maker of Unix servers that most pundits thought was long past its prime.
Of course, now that Oracle owns the patents and copyrights related to Java (a technology originally created by Sun), they have gone after Google with a greedy vengeance, chasing the revenue potential to be found in Google’s use of Java in Android. With the real possibility that Oracle could win billions, both in terms of immediate penalties and from ongoing fees derived from the growing success of the Android platform, $7.4 billion would have been a bargain for Google.
Oracle’s push to derive revenue from Android isn’t the only reason Google should have bought Sun. Rather, it was a strategic error that deprives Google of a standard software development platform at a time when the platform landscape is in flux, shaken in recent years by a device revolution that has driven Apple - and Android - to new heights.
To be fair, mergers have a long history of going horribly wrong. Microsoft is often the poster child for this, which was part of the reason I opposed Microsoft’s attempted purchase of Yahoo. The WebTV purchase eventually led to Microsoft IPTV / Mediaroom, though only if you think “lead” means to take 8+ years to get to a point where you have a viable IPTV platform built around technology that was at right angles to other TV and video-related efforts at the company. The textbook case of a failed merger, however, is surely Microsoft’s $500 million purchase of Danger, a company co-founded by Andy Rubin, the guy who now manages Android at Google. Danger’s resources were focused entirely on the ill-fated KIN, a device that was a distraction at a time when Microsoft should have had a laser focus on Windows Phone 7.
Mergers don’t always run off the rails. Google’s purchase of YouTube for $1 billion can be labeled a success. Granted, bandwidth costs make it very hard for Google to make a profit from YouTube, but YouTube’s status as the de facto source of Internet video has surely given Google opportunities in other areas. Access to YouTube is an essential feature in both IP-enabled televisions and smartphones.
And, though Stephen J. Vaughn surely disagrees, I think Skype could prove a good purchase. It’s hard to build a brand that people associate with voice and video communications as strong as Skype, and the fact that the Skype protocol is strange matters to end users about as much as the twists in the plumbing snaking through their walls. I think Skype, potentially, could offer the same advantages to Microsoft that YouTube offers to Google, at least in the realm of audio and video communications. Of course, between potentiality and actuality there are chasms filled with rocks and alligators, so we’ll see how things proceed.
A Google purchase of Sun Microsystems would have served as a big boost to Google’s software platform ambitions. Solaris probably mattered little to Google, and perhaps could have been spun off to a company who cared more about custom hardware (Oracle being a prime candidate). Sun’s software was what really mattered, and the software Sun created or owned would have given them a strong development API that rivals .NET or Objective-C / Cocoa that was fully owned by them, as well have provided a strong office productivity suite that could be augmented to complement their online word processing tools in unique ways. Those two aspects would be real weapons against Microsoft, striking as they did at the core of the Microsoft revenue machine (the Windows platform, and Office).
Had Google owned Java, Google could have given it a special place as an API, helping to boost popularity of an already popular technology. Most programmers emerging from computer science departments have strong familiarity with the platform, and Java has a strong presence in multiple industry segments, from servers through televisions and mobile devices. Google certainly wouldn’t cease to support other development technologies, or stop pushing web interfaces as the way to make client applications. However, even on the client, there is still much value to be derived from locally-resident applications. Andy Rubin and the Android team understood that, which is why they chose Java as a development technology and not a wrapper around HTML / CSS / Javascript, as Palm did with Web OS.
The timing for such a positioning couldn’t have been better. Microsoft is currently on the back foot in both phones and tablets. The platforms that have come to define the future of computing aren’t emerging from Microsoft labs.
Further, Microsoft is preparing for a major evolution in their UI strategy that has put strain on their normally strong relations with developers. Microsoft’s recent unveiling of concepts underpinning its new Windows 8 excited a lot of people. The opening screen will support applications written in HTML, CSS and Javascript, which is quite interesting and an important step forward. Windows developers, however, noticed the omission of any mention of .NET, creating the impression that .NET applications wouldn’t be first-class citizens on the new Windows 8 desktop. Incomprehensibly, Microsoft refuses to confirm or deny whether this is, in fact, the case.
Perhaps, as many have noted, this is all a colossal bit of miscommunication. Yet, Microsoft is normally quite good at developer communications, making this “misstep” out of character, lending credence to the notion that they are seriously contemplating orphaning aspect of their .NET client story. As Tim Anderson at The Register recently stated, “From the outside, it still looks as if Microsoft’s Server and Tools division is pulling one way, and the Windows team the other.” As a former Microsoft employee, I always found it astounding how poorly different product teams cooperated with each other. The fears of the .NET development community seems well placed.
Microsoft’s pain, however, could have been Google’s gain. Internecine struggles at its Redmond-based competitor would present a golden opportunity, had Google owned a platform they could present as an alternative.
Google wants to be the platform for the Internet age. Its focus on browsers is part of that, but clearly, the success of Android moves the focus even more firmly in Google’s direction. Every platform company needs a standard API. Apple has Objective-C and Cocoa. Microsoft has .NET, even if they have a hard time forcing internal teams to hew the line on that. Google could have had Java. Given the generally favorable impression the open source community has of Google, they could have directed development efforts atop Java in directions useful to Google.
A similar analysis applies to OpenOffice, an open source word processing suite that Sun bought in 1999 under the StarOffice name.
To be frank, OpenOffice hasn’t made much of a dent in the dominance of Microsoft Office. That, however, doesn’t mean it isn’t a useful tool to help to erode the importance of a product that is a source of a large share of Microsoft revenue. Owning the copyrights to Open Office would give them the power to guide its development to align with Google’s cloud-based document editing tools. It would give them a strong suite of offline business productivity tools, and as with Java, Google would likely have had considerable success in steering the community down a path that aligns well with Google’s cloud strategy.
Google is certainly not going to fall over from any settlement, forced or otherwise, with Oracle. However, the fact that the owner of Java is an antagonist makes it less likely Google will standardize in any real way on Java. That’s a shame, as Java is good technology that would benefit from a big company like Google backing it with real money.
It took nearly a decade for .NET to entrench itself in any real way as the standard platform for Windows development…and even then, it still hasn’t completely displaced traditional WIN32. Objective-C and Cocoa has been around even longer by way of NextStep. It takes time, in other words, to establish a platform. Quite simply, Google won’t have time to make a real alternative on its own. Their only chance was with Java, and that chance was missed.
