What Would the Operating System Look Like if It Were Designed Today? The Chrome OS Q&A
“I think the real question is (that) if you were going to design an OS today, what would it look like? The OS that we’re using today is kind of in the model of a ’70s or ’80s vintage workstation. It was designed for a LAN, it’s got this great display, and a mouse, and all this stuff, but it’s not inherently designed for the Internet. The Internet is this resource in the back end that you can design things to take advantage of. You can use it to synchronize stuff, and communicate stuff amongst these devices at the edge.
A student today or a web startup, they don’t actually start at the desktop. They start at the web, they start building web solutions, and immediately deploy that to a browser. So from that perspective, what programming models can I give these folks that they can extend that functionality out to the edge? In the cases where they want mobility, where they want a rich dynamic experience as a piece of their solution, how can I make it incremental for them to extend those things, as opposed to learning the desktop world from scratch?
” – Ray Ozzie, 3/10/2008, in an interview
with Om Malik
Of the two things I’ve consistently said about Google since, oh, 2004
or so, one is still true: Google is still authoring aggressively operating system independent services. Today’s announcement of a desktop operating system based on Chrome
What is less obvious is whether or not Google’s to date successful approach of not trying to out-Microsoft Microsoft is still in play, or whether they’re finally succumbing to the blinding temptation to own, well, everything. While I could build and defend either case, the Department of Justice is almost certainly going to concern itself increasingly
with the latter scenario.
More immediately, the pending Chrome OS is in many respects the realization of a decades old vision, as Rafe reminds us
. Known by a variety of names, the most common appelation of Network Computer
would be easy to apply to a Chrome OS equipped node. It’s not quite a network computer, of course, in that the netbooks that will be the Chrome OS’s first hardware platform typically have at least some on board disk space. But the basic elements are there; a hardware platform and operating system pairing that is optimized to do one thing superbly – access network services.
As much as Oracle and Sun’s – isn’t that quite the coincidence, incidentally – vision for network computers failed, it appears in hindsight to be a classic case of being early to market. Remember what the world was like in 1996? No cellular data networks? Hell, the wifi patents were still in the filing process that point. And I won’t even get into what the web looked like back then; fortunately the crimes I myself committed against it are lost behind a prescient robots.txt file.
Even with more ubiquitous connectivity and an accelerating web, however, the outlook for – and implications of – the Chrome OS depend on a number of variables. To explore these, let’s turn to our old friend the Q&A.
: Before we begin, do you have anything to disclose?
: A variety of companies that are involved in this discussion are RedMonk clients, from Adobe, IBM, Microsoft and Sun to Canonical and Red Hat. Google is not a RedMonk client; RedMonk, however, is a Google customer.
: Ok, for those that missed it, can you summarize the news?
: Certainly. Last night, via their official blog
, Google announced the Chrome OS, “an open source, lightweight operating system that will initially be targeted at netbooks.” Essentially it marries Google’s Chrome browser to a Linux kernel to produce a basic, browser based operating system. The windowing system will reportedly be new as well, indicating that they’re eschewing traditional Linux desktop environments like GNOME and KDE in favor of a potentially from scratch approach that, like Chrome the browser, is web front and center. I’m very curious as to what they’ve got planned for the windowing system, actually.
: Is the project open source?
: Not as yet, but they’ve promised that it will be. In the meantime, as has become typical with Google projects for better or for worse, they’ll work on it behind the scenes before dropping it on Google Code.
: Is the fact that it is – or at least will be – open source important?
: Absolutely. For Google, because it may compel wider adoption, testing, and development, but also for open source competitors, who may – depending on the licensing – be able to borrow liberally from the project.
: What hardware platforms are supported?
: At present, the goal seems to be x86 and ARM, which would offer the operating system the ability to penetrate the overwhelming majority of the netbook market, not to mention expand into laptops and desktops or smartphones.
: Doesn’t Google already have an operating system for smartphones in Android?
: Indeed they do, but Chrome OS is a very different animal from Android. While some are raising legitimate questions about the intelligence and viability of maintaining either two operating systems
or so many open source projects
, this to me is a more logical step for Google than would be attempting to push Android onto the same platforms.
: Why? Why didn’t they just double down on Android on the desktop? Or, alternately, skip it in favor of Chrome OS?
: At present, phones and desktops are materially and necessarily distinct in their relative strengths and weaknesses, and by extension, the application’s ability to leverage or minimize same. Android, which is essentially Java married to a Linux kernel, is designed expressly for handsets. The Palm Pre may yet prove that a web only model can compete effectively against the richness of the iPhone, but Android would indicate that even Google felt compelled to turn to native code in an effort to match the Apple experience.
On the other hand, what do people use netbooks for, typically? Checking email? Browsing the web? Instant messaging? Facebook? Paying bills online? Why, then, would you push the Android rock up the hill, when a simple browser based interface would more than suffice for the majority of needs? Exactly. You wouldn’t, and Google didn’t. I never believed
that Android on laptops made sense, personally, because of the applications question.
: What’s the applications question?
: Platform success is nearly always achieved via application volume. Platforms that attract volume tend to succeed, those that don’t, don’t. While there are a great many variables at work, application volume is typically the peformance metric.
Which always made Android a questionable choice for me. Forget the fact that a world of rich applications isn’t exactly in Google’s best interests (contrast that with the Chrome OS experience) and just ask yourself this: is it easier, as a developer, to develop for the web or for a Java-like platform in Dalvik? The answer is, of course, the former. So that was always strike one versus Android. Strike two was the fact that virtually every other competing desktop platform – Linux, Mac and Windows – would have a superior application volume from day one, and the ability to indefinitely sustain that by running Android apps themselves along with other choices.
Strike three? It’s more difficult for Google to advertise in – and thus monetize – Android rich clients than it is their own web applications, which are presumably intended to be at the heart of the Chrome OS experience.
Add it all up and Android was never a good choice for the netbook form factor; whether it remains a quality choice on the handset depends – you guessed it – on application volume.
: Is the idea of a browser based operating system new?
: Not really. The GNOME crew, for example, has been discussing the possibility of an “online desktop
” for a few years now, and even if it’s not what Chrome OS envisions, the concepts are similar. Microsoft, a rich client advocate if ever there was one, even has its own browser project
in Gazelle that encompasses some operating system like features. Good OS has been tinkering with something very similar to Chrome OS in Cloud
But more to the point, the netbook market itself is almost a browser based operating system. While netbooks today are more often than not shipping with a version of Windows XP, it was Linux interfaces that actually pioneered the market. And to talk to the early users of devices like the Eee, they were essentially jumped up web terminals, since few had the interest, inclination or ability to explore the application catalog for a heavily customized Linux desktop.
We know, in other words, that there’s a market. How big that market is is one interesting question. How big it could get as network delivered browser based applications get better is another.
: What about Gartner’s research
that indicates that users are underwhelmed by SaaS applications? If the affection for network delivered application wanes, won’t that negatively impact the Chrome OS?
: An observation: the audience here is important. Gartner reports the survey subjects as “users and prospects of SaaS solutions in 333 enterprises in the U.S. and the U.K.” Enterprises, in other words. Which are not only Gartner’s typical clients, but usually the last to know
when it comes to technology. Enterprises may or may not like SaaS as an application delivery approach, but it won’t matter because an increasing percentage of their choices will be delivered that way regardless.
Besides, enterprise buyers are not the target market for netbook sales anyhow. The market, which was essentially created overnight under the noses of the big software and hardware vendors alike, is overwhelmingly consumer focused. And apart from hardware specific requirements like iPod/iPhone compatibility or gaming, their typical computer usage – Facebook, Gmail, online banking and shopping, etc – is increasingly operating system independent.
Just as Google always intended.
Eventually, just as we’ve seen with cell phones, instant messaging and webmail, netbooks may make their way back into the enterprise. But until then, they should have a healthy market in front of them, enterprise appetite or no.
: Speaking of iPods and such, what will the device compatibility story be?
: Initially? Probably zero, with the exception of maybe printing, which is more seamless now, typically, on Linux than it is on Windows. No, I don’t expect Google to tackle the very difficult question of device compatibility
at all in the initial releases of Chrome OS, if only because it’s a slippery slope.
This will be, I’m sure, one of the defenses that Apple and Windows employ against the potential threat of Chrome OS. And for many consumers, it will be a compelling one.
: Why do you think Google decided to create (yet) another Linux distribution, rather than embracing popular desktop and netbook options like Ubuntu’s?
: Not having spoken with them, I can only speculate, but my guess is very simple: speed. Besides the relevancy of its results, what’s the defining characteristic of its search engine? Speed.
Likewise, when Google broke with Firefox via Chrome (the browser), my view
was that they felt that a new, from scratch browser was the only means of competing effectively with one of the native clients’ last remaining advantages: speed. This has been largely born out in their development; while the Tracemonkey enabled Firefox 3.5 is clearly superior to its 3.0 parent, Chromium for Linux is one of the fastest applications I’ve seen, period. Browser or no. It hasn’t been enough to compel me to switch, as yet, but the emphasis on speed and quickness has paid dividends for them.
So while I have no inside information on the subject, I suspect much the same rationale was at work in the design and development of the Chrome OS. Linux, in spite of some dedicated efforts, has yet to introduce device like low latency into the desktop. Booting into a modern desktop with all of its bells and whistles means overhead. By eliminating the traditional operating system overhead in favor of a simple, high peformance browser, boot time and general performance should be improved dramatically.
And that matters. Speed is a feature
: How much does it matter?
: Enough that I briefly entertained the idea of cutting over from Ubuntu – which boots on my SSD equipped X301 in a bit over 20 seconds – to Moblin, available in under 10. I didn’t make the switch, obviously, because I wasn’t quite ready to lose my application portfolio in the process, but it matters.
And yes, that’s even in a world in which suspend works flawlessly. Consider that the feature that Apple chose to highlight in its latest iteration of the iPhone – the 3GS – was speed. If that still doesn’t convince you, think of it this way: if a consumer ever picks up, or sees someone else pick up a small computer that boots in a few seconds, that will in all likelihood be something they’ve never seen before. Sad, but true.
: What about the question people like Gordon Haff are asking
: why would the “Google Chrome OS would have mainstream impact when desktop Linux has not?”
: Because while the Chrome OS is Linux, it’s pretty significantly differentiated. Linux, we all need to remember, is the kernel; the things that a user will see and interact with are something else entirely.
First, there’s the speed. I’m guessing that over time Chrome OS will be pretty close to instant-on, with operational latency as impressive as that Google’s achieved in the web world. While mainstream Linux distributions have made many improvements in desktop performance the last few years, it won’t match that soon, if only because it’s handicapped by having to run more than just a browser.
But the single greatest differentiation might just be that: the browser as the interface. Absent, mostly, is the need to learn the differences of a new operating system; your interface is “just” a browser. While that imposes some significant limitations, clearly, it also dramatically lowers the barriers to entry to the product. I know a great many non-technical people that have picked up Google Chrome with with no introduction; I can’t say the same of any currently available operating system, Mac included.
Last, there’s the power of brand, which observers will discount at their own peril. Ask Apple about that sometime.
: So you’re bullish about the distribution?
: Not as much as I recognize the opportunity it’s targeting. Frankly, I have mixed feelings about the distribution. While I personally would love to have a capable platform that booted in a few seconds and welcome the kind of back-to-the-basics innovation that Chrome OS represents, I’m not particularly eager to see the introduction of (yet) another distribution into an already fragmented market. Could Chrome OS be good competition for Ubuntu Netbook Remix, Moblin et al? Sure. But it could also introduce significant educational challenges as a variety of distributions with a variety of interfaces jostle for marketshare.
All of that said, I do appreciate this kind of innovation.
: Which kind is that?
: The “we don’t have to do things this way just because that’s the way they’ve always been done.” Much like Wave
rethinks collaboration, Chrome OS is a reconsideration of what an operating system should look like in a heavily networked environment.
I know there are those that question Google’s innovation
track record, and both Chrome OS and Wave could certainly go the way of the dinosaur, but I appreciate that someone is asking these questions.
: Back to marketshare, do you think this was targeted at Microsoft Windows?
: If Google’s smart, and I think generally they are, they’re not targeting a company but opportunities. And there’s no question that there is opportunity for innovation in the operating system market – significant opportunity. Consider that Google’s seemingly snarky assertion – that “the operating systems that browsers run on were designed in an era where there was no web” – is not only true, but confirmed by none other than Microsoft’s Ray Ozzie (see the quote above).
: What impact do you expect this to have on the folks from Redmond’s operating system?
: Well, at the very least I think it’s a wee bit early to be writing the obituary
for the Windows operating sytem, as some are doing. Chrome OS represents a significant problem for the folks from Redmond, to be sure, but – like OS X – as long as there are applications and devices that exclusively operate on the Windows platform, they’ve got adequate insulation at least at the high ends of the market. For the short term, anyhow.
Over the longer term, of course, the transition to web applications that Chrome OS is specifically designed to leverage and accelerate represents a serious threat to Microsoft’s primary revenue sources, but that’s not news. It wasn’t even news five years ago. So while Microsoft is probably no happier about this news than Apple – or Canonical et al, for that matter – Chrome OS is not the death knell for the existing products any more than Chrome the browser killed Firefox.
If anyone is threatened by this, actually, it’s probably the Linux distributions.
: Why is that?
: The various Linux distributions are less highly differentiated from Chrome OS than is Windows – or, should they decide to enter the market, Apple – ergo they will likely be the most impacted by its arrival. Customers purchasing Windows are typically doing for specific reasons; they rely on Windows compatible applications, they’re used to it, and so on. The competing Linux distributions enjoy no such visibility yet, however.