Original URL: http://www.theregister.co.uk/2008/09/04/google_chrome_analysis/

History shaped Google's Trojan Horse

Unravelling the Chrome masterplan... with Windows 2.03

By Andrew Orlowski

Posted in Hardware, 4th September 2008 13:26 GMT

Analysis When people buy software - buy it in seriously large amounts - it isn't just today's binary they're choosing. They're buying what they think is a bit of the future - they're buying a piece of risk insurance. This explains why very mature and well-proven systems often lose out to the Newest Kid on the Block. It also explains the enduring effectiveness of FUD and Vapourware.

And it's not just software. From TP monitors, to minicomputers, to Novell Netware, recent history is full of examples of perfectly splendid systems being thrown out and replaced with something that doesn't live up to the billing - and perhaps never will. Which sounds wacky, but that choice is being made on the rational calculation that the software or hardware of choice today won't be made or supported, or the standards that bind the parts of the system together will become obsolete. (Which leads to the same thing.)

Sometimes a brave company bucks the trend. Most famously Microsoft refused to "eat its own dog food", and stood firm against the move to client/server computing running PC or Unix-based databases like Microsoft SQL Server, instead insisting that its mission-critical accounts department ran on, er, an IBM AS/400 mini.

But by and large, the strategy works very well for companies that trumpet a "paradigm shift", or "new era in computing", and convince people that they own a secret part of the future - one that no one else can yet see. It worked for Microsoft, and Google hopes it will work for it, too. The Chrome browser today is little more than a piece of demoware, but it's not just about "today", is it?

Before we see what Google is hoping to achieve with Chrome, let's take a look at a precedent from history that I find quite spooky. Old-timers may excuse this brief wallow in nostalgia.

In the 1980s, PC business software was dominated by three names. Ashton Tate, Lotus and the WordPerfect Corp. The former two produced dBase III and Lotus 1-2-3, which were practically mandatory. Each product had what your modern, New Age marketing-droid would call an "ecosystem" around it - the value of the choice was as much in third-party add-ons and libraries of macros and scripts, as in what came out of the box. Developers skilled in these black arts were plentiful too.

For their part, Ashton Tate and Lotus had grown fat and lazy from astronomical growth, and had been slow in updating the software. They had seen off competition from integrated suites, and looked formidable enough to keep superior rivals from gaining much market share. And they were very expensive - dBase IV retailed for $795 in 1990.

But buyers, who were in no rush to migrate, knew there were two events in the coming years that might force them to re-evaluate at some point. 32bit computing would eventually supplant the limited address space of DOS running on 286 or 8086 machines, and eventually - at some far off date in the future - graphical user interfaces would come to the PC.

Microsoft knew this too, but it had a few problems. Its own clunky GUI, Windows, offered no advantages to the business user. The giants of the DOS world wouldn't run very well inside Windows - if they ran at all. There was no unique killer application for Windows, either.

Worst of all, few people really believed that Microsoft owned that vital secret of the future, or knew something no else knew. Apart from DOS, Microsoft simply sold a few compilers, while its own applications rarely got to a medal position in the shoot-out comparison tables in the computer press. And that was about it.

As Software Magazine, reviewing Windows 2 in 1988, wrote -

"There are challengers, including Desqview, and entries from Hewlett-Packard, Xerox and IBM's own Presentation Manager."

In other words, if the PC went GUI, it would probably be thanks to one of the grown-ups. Apple had priced itself out of the business market and refused to license the software. However, what credibility Microsoft had rested entirely on hanging onto IBM's coat tails with its work on OS/2. And few people had any great enthusiasm for a future that returned control of the industry back to IBM.

So Microsoft bundled a Windows "runtime" with one of the few prestige applications that had been ported to Windows. Typically this might be Aldus Pagemaker, or Microsoft's own Excel spreadsheet, because there weren't really any other heavyweight Windows applications. The runtime was a limited version of Windows that started when PageMaker ran, and ended when you closed it.

Unravelling history

Like Microsoft 20 years ago, Google wants to shift users to a new platform - its own - for which there is much hype but no great enthusiasm. Like similar migrations the new platform offers very few advantages - and plenty of disadvantages. Not only are great chunks of functionality missing, but even when you're supposed to be "online and always connected", you might not be.

There have been plenty of hiccups in the "cloud", recently. As Ted Dziuba wrote here recently, it's captivated the investors for several dubious reasons - one of which is that a "cloud" is ever so easy to draw on a White Board.

That's where the "runtime" comes in. Today, Chrome is simply a technology demonstration - and I can't see Firefox users with their carefully-cultivated selection of add-ons, or Opera users, making the jump any time soon. But Chrome is a Trojan Horse for bundling Google's Gears onto your PC - and in the hope that manufacturers look to Google services for new Eee-type lightweight PCs, perhaps running something like gOS, the Ubuntu-derivative.

Gears is simply designed to make Google's online services more attractive, and makes it looks like Google's is setting the standard: leading where everyone else follows. (That isn't entirely unfair.) And as a technology demonstration, Chrome succeeds.

Google has two powerful arguments for software as a (Google) service: it may be cheaper than licensing Office, and less complex than running client/server in your office. Uncannily, that's the two things that helped swing the market Microsoft's way, too. Migrating to a GUI required more powerful and expensive PCs, but it would save large amounts in training costs.

Microsoft then played its ace: it began bundling a not-very-integrated "suite" of applications for around $500 - less than the price of dBase or 1-2-3, and that's before you'd bought the essentially companion software such as Clipper. In a recession, that started to look quite attractive.

Windows also offered a "cockroach" alternative to some of the grander vapourware designs on offer. Rather than wait for the Next Big Paradigm Shift (there were many of these vision-things being touted back in 1990, invariably including the words "Object" and "Architecture"), users smuggled in a copy of Windows for Workgroups and tried to get it running on the company network. Google services have the same appeal: people simply start using them.

So will Google succeed? Well, you tell me (below). But your 80s throwback will offer a couple of perspectives, that I've looked for this week, but failed to find.

Clever is not clairvoyant

One of these is that "the company that knows secrets about the future™" is a myth created by the press - particularly the glossy end of the US business press. It's a powerful narrative, and suits their lazy writers, but the reality turns out to be very different.

Years later, we discover the company was simply blundering on in a state of chaos, slapping tactics together until they passed for a strategy, and winging it. And so a consequence of this myth-making is that it makes the poster child - the Google or the Microsoft - look much cleverer and more coherent than it really is. It's an elaborate game of bluff.

At the time Windows offered business a cheap and cheerful "standard", but Microsoft's success was not based on technical excellence - on any unique knowledge of 32-bit computing or excellence in UI design - but rather more to its iron grip on the PC distribution channel. PC manufacturers paid Microsoft whether they shipped MS-DOS and Windows with the PC or not. So why ship anything else? Antipathy to IBM helped Microsoft enormously, too, of course.

But I simply don't see where Google has the same grip over routes to market that Microsoft could exploit. And while costs can certainly be lowered by throwing away all your useful software, I don't see that Microsoft generates the same animosity that IBM once did. I'm confident that we'll be using web services more, as they get richer and more functional.

I'll predict that Linux will thrive as a kind of bootloader on low-end PCs designed to use these services. And that Microsoft, as a result, will face continual margin pressure on Office and Windows in the years ahead. But I can't see either Microsoft, or the idea of local applications, fading very far from view.

While much of the press has creamed itself over Chrome this week, it's almost rude to point this next one out. When there's one computer serving the planet - even if it's Google's - that's a single point of failure.

And in that sense, Google's vision of computing looks less like a piece of risk insurance, than a very big risk indeed. ®