Chrome Detailing: Previewing Google's New Operating System - HotHardware
Chrome Detailing: Previewing Google's New Operating System

Chrome Detailing: Previewing Google's New Operating System

Last week, Google open-sourced its Chromium OS project, more than a year before the operating system is scheduled for release. In doing so, Google hopes a variety of developers and companies will become involved in the project, and has pledged to release regular updates as well as a comprehensive log of bug reports and fixes. We've spent a few days playing with Chromium and exploring its design and come away impressed—but not without questions.

This article should not be considered a review of Chromium in any way—at most, it's a very early preview. We won't be discussing performance or spending too much time on the operating system's "look and feel." Google has stated up front that a number of UI changes are in the pipeline.

What we are going to talk about is Google's design vision for Chromium, the unique benefits it offers, and a bit of why Google is throwing its hat into this particular ring in the first place. Chromium, after all, is a Linux-based OS entering the smartbook/netbook market at a time when said product segment is already being served by a variety of Linux distros (some customized to the hardware, some not), XP, and Windows 7. In the midst of all these options, do we need another operating system?

We just might.

Understanding Chromium

Google has put together a 3.5 minute video that does a pretty good job of explaining Chrome's design philosophy. For those of you who don't feel like watching, we'll discuss it after the jump.

Chromium is designed around the premise that computers have largely evolved into boxes from which we access the Internet. If you've ever lost Internet service (but not electricity) for a protracted period of time, you've probably seen the company's point. Evaluated in terms of what a modern system is capable of doing, getting online is but one, minor feature. From a user perspective, however, Internet connectivity is a matter of tremendous importance—lose the net, and your system transforms from a communications platform/multimedia hub into a box for writing papers, gaming alone... or something.

Google's view, although the company doesn't say it in so many words, is that the lot of us—Linux, Mac, and Microsoft users alike—are using a device initialization and access platform designed nearly thirty years ago to access the cutting-edge capability of today. Imagine, for a moment, that we were shipping 1080P content on VCR tapes at 50 tapes per movie, and you get the picture.

On the left, old and busted. On the right, New Hotness.
If computers are basically Internet portals and almost everything we do takes place online, then why not cut out the middle man altogether? That's where Chromium comes in. The OS is very clearly based on the Chrome browser, but Google has already begun implementing the more advanced UI requirements that an OS requires.

Wherever You Go, Here You Are

Chromium simplifies the concept of an application the same way it simplifies the operating system i.e., by throwing most of it away. Chromium apps won't just interface with the web or contain web-based components, they'll exist there. The 'cloud' in cloud computing thus transforms from an abstract concept of computing resources or storage floating vaguely in space into something much more unique and personal. The cloud has literally become your cloud—your life, online. The OS is tied tightly enough to the cloud that online connectivity is vital in order to practically operate or make use of the device.

Google's version of a task manager.  Note the amazing amount of goat teleportation...
We moved a herd last night (rounded up to the nearest full goat.)

All this focus on Internet connectivity and web applications may have you wondering what sort of local storage options are available, and the answer is currently up in the air. Several of Google's presentations imply that the system has no internal storage at all (at least none that is user-accessible). At other points the company states that no hard drives will be supported, but that systems will include varying amounts of flash memory.

To summarize the sales pitch: Chromium gets rid of all the crusty old legacy garbage, moves storage online where it's both universally available and backed up, provides a platform that finally integrates browser and OS, all while providing a fabulous, multimedia-rich online experience. Everything you currently do offline will be available online, seamlessly provided by a content platform that presents a universal, standards-based framework rather than a hodgepodge of browsers, security bugs, and broken standards support.

Chrome's memory manager

Rather than attempt to pass verdict on an OS that isn't out of beta, we've devoted this section to a few of Chromium's potential greatest strengths and accompanying challenges. If you've got a different list than we do, feel free to sound off below in the comments.

The Potential
  • A Genuine, Unified Windows Alternative: Linux holds part of the netbook market, but the majority of the space, particularly the higher-end models, have opted for a Microsoft operating system. Chromium is Linux-based and it'll be backed by the financial and technical assets of Google. If Google woos developers appropriately and develops the proper toolkits, Microsoft could find itself with a true head-to-head competitor, the likes of which it hasn't had to deal with in decades. Google, like Microsoft, can afford to take its time and get it right, even if it takes a few years to do so.
  • Competitive non-x86 Products: The overwhelming majority of netbooks currently in existence are based on Intel's Atom and powered by an Intel chipset. There's nothing wrong with that—the point is not criticise Atom—but a new, processor-agnostic OS could drive innovation in both processor and platform design across the industry. Currently, an x86 processor has a massive inherent advantage because it runs Windows—and so does everything else. Over a period of time, there's no reason Google couldn't simultaneously invest in whatever software tools are needed to ease the transition while providing OEMs with the information they need to make deploying a non-Windows product as easy as a Windows one. Indeed, the cost of 'transition' is designed to be nonexistent. If you're familiar with a browser window on Windows, you're familiar with a browser window in Chromium.

  • All Hail Our Corporate, Web 2.0 Overlord: This last advantage is more a gain for Google than it is for your average consumer, but it's a gain—and a corporate goal—worth talking about. It's no secret that Google wants to be more than just your search engine; if Chromium takes market share, it, and by extension Google, will be front-and-center of your entire Internet experience. Again, this isn't meant to be inherently evil-sounding, but if it works, Google's brand presence would expand immensely.
The Challenges
  • Making It "Just Work": If the browser is the operating system, the browser window has to be able to handle a lot of different types of content, and it needs to handle it well. Google has said it intends to use HTML5 to pull off quite a few of its planned features, but it'll need other capabilities, like direct access to the video card. From a technical perspective, there's a lot of things that need to happen under the hood in order for Chromium to present users with a smooth, stable, experience. The company also has to have a plan for what happens when something doesn't work.

    Once upon a time, if a program or system driver crashed, it took your OS with it nine times out of ten. Then XP came along, and an application blowout didn't necessarily mean an automatic reboot. Vista and Windows 7 have further advanced the concept—if your display driver crashes these days, Windows will often freeze briefly, then resume normal operation (with a note of what occurred). If you kill an instance of Firefox within Task Manager, the app remembers the websites you had open, and offers to open them again. Chrome, the browser, offers this sort of functionality, but an OS needs layers of robustness that a browser can take for granted.

  • Offline Functionality and Local Storage: Google can talk all it wants about how the web has become synonymous with computing, and it sounds accurate until you stop and think about all the exceptions. These include, but are by no means limited to: Areas without free WiFi, areas with theoretically free WiFi with practically zero functionality. Hospitals, airplanes (without paying exorbitant costs), odd corners of the house, power outages, road trips, and pretty much the entire Louisiana Purchase.

    Cellular cards could solve some of these problems but introduce others, especially if you don't live in 3G territory. To date, Chromium is being discussed as a web-centric operating system designed to put you in direct, lightning-fast contact with the content you want most. That's all to the good and it gives the company the chance to talk about what really sets its OS apart from those dull desktop types. A year from now, a lot of Chromium's appeal could be riding on what users can do with one when they aren't connected to the Internet or want to save content locally. There are occasions, after all, where 'the cloud' is the very last place you want certain information to reside.
One final bit of irony. Over ten years ago, Microsoft was sued for bundling a browser with Windows 98. Does it amuse anyone else that Google is bundling an operating system along with their browser?
+ -

I look forward to giving it a test drive when it's ready for prime time. The concept is a good one but I think that most of us are deeply entrenched in the practice/concept that we have to store our stuff here at home. There is a whole lot of data here at home too. My at home storage capability totals almost 7 terabytes, and I am actually using 3.4 TB of it at this time. What would such data requirements cost to store on Google's servers?

The idea of everything being on the web and not here, in our physical control, is gonna take some getting used to. Web based security for all of our files will be an issue for allot of us too. Almost every online entity has been hacked to some degree in the past and we know it. So people may never trust the 'Cloud' as much as home based mega-sized storage.

I'm still interested in checking it out,,,,,,,,,,,maybe installed onto a little NetBook for use when I go on my road trips.

+ -

<edit> Nevermind, it appears I need more coffee prior to doing mental math at 2 in the afternoon... <sigh></edit>

+ -

This might appeal to users that generally wouldn't use Linux, because Google will simplify things even more than distros like Mandriva and Ubuntu already have.  For people already using Linux:  I really don't think there's much appeal to running this, vs. running a full Linux distro with Google's apps on top of it.

The interesting question is, are they going to settle for this to be just another Linux distro, or are they going to address the real reason people stick with Windows over Linux:  Where are the games?

 Are we supposed to only play web-based Flash and HTML5 games?  For all my whining regarding Microsoft's locking in of developers with their proprietary DirectX APIs - it has been an extremely effective and foresighted defense for their monopoly.

If Google seriously wants to dent MS's market share, they need to create their own superset package of OpenGL/OpenAL/SDL, smooth out all of the rough edges, and get the developers on board.  And, if they could do it all in the Mono framework so that the same apps can run unchanged on Linux, under Windows, and even on other CPU architectures - all the better for user choice.

+ -

I agree that if Linux 'Gave Good Game', they would be 'In The Game' more.

The Ubuntu And Linux Mint Gloria distributions that I've tried out are excellent replacements for Windows and have full capabilities, except for the kind of games that Windows excels at.

Linux would take a huge bite out of Microsoft's bloated, expensive ass if it could run all of those games,........

+ -

This is why I think the Wine project (and by extension CrossOver) is so important.  And no, not just the one where I sit here and drink wine throughout my vacation.

It's the one tech that allows Linux users to still run a lot of games and those few critical Windows apps that they just can't find a Linux equivalent (ex.  DreamWeaver) for - or for which they need 100% compatibility (MS Office vs. OpenOffice).

+ -

So, just a browser is expected to:

- Connect printers and other USB flash drives without any custom downloads

- Access to the webcams

- Import pictures from my camera

- Import videos from the handycam

- Burn DVDs with music/pictures etc

- Make calls (speakers/microphone)

- Access my mobile phone

I see a long path before an average user can use Chrome, by which time the market for netbooks would be pretty much consolidated.

My 2 cents.

+ -

Remember:  ChromeOS is just Linux with a new Xserver/desktop environment:

- Connect printers and other USB flash drives without any custom downloads

In Linux, you can configure printers via the webbrowser - CUPS has a web interface.  I've yet to see a USB flash drive that wasn't automatically recognized when connected.

- Access to the webcams

HTML5 supports this.  Adobe Flash already does this inside the browser too.

- Import pictures from my camera

USB devices are recognized by the kernel.  You can access them from the webbrowser as simple as uploading pics to Facebook.  HTML5 also supports the concept of local storage.

- Import videos from the handycam

Same as above.

- Burn DVDs with music/pictures etc

Linux follows the Unix philosophy where there are many utilities that do one simple task well, and support piped interactions with everything else.  Most CD recording software on Linux are just Qt/GTK GUIs that sit on top of the command line utilities mkisofs and cdrecord.

A web interface could do the same thing.  I'm literally talking an hour or two for a PHP hacker to get something working, if someone hasn't already written one already.

BTW:  most people see "command line utility" and think DOS.  It's not like that at all:  It's a true 64-bit recording engine.  Abstracting the interface from the engines allows both to be developed and bug-fixed independently.  I.e. a fix in cdrecord improves every single CD recording utility in Linux.  Also, it's this abstraction and specialization of function that allows web interfaces and other new GUIs to be built so easily in the *nix world.

- Make calls (speakers/microphone)

HTML5 already supports this, as does Flash.

- Access my mobile phone

It's just a USB device... I plug my phone into my Linux box and it shows up as a drive automagically.  With a driver specific for your device, you could even do address book synchronization, etc.

+ -

There's a USB sharing NDIS driver for windows that makes windows mobile internet connection sharing like built into windows mobile 6.x phones work in Linux. It does require a little command line work to get it working the first time. and SynCE is another option.

+ -

Google have said they see both ChromeOS and Android as "flanking technologies", meaning these are not intended to replace general purpose OSs outright. Neither ChromeOS nor litl's litlOS are in any sense general purpose at this time. The general purpose OSs like Windows and OSX will continue to be used for what they are good at, but these come with the usual maintenance and bloatware hassles for Joe User not to mention heavy resource use.

When most people do their stuff on the web, a web-oriented OS is the way to go. The difference between ChromeOS and litlOS is that the latter is already on the market and is designed specifically as an appliance OS for the litl webbook - so we already know exactly what it is being be used for and in what ways it can grow. Litl is focused on making leisure use at home easier for all the family and on making the experience of streaming media and other content even better through litl channels. Since we have a remote update system that upgrades channels and software while you sleep, our software will continue to grow on our device without any intervention on the part of the user.

+ -


Take the word "support," and add "Just as fast, just as reliably, and just as well." That's really the question here. And in reality, Chrome actually needs to do it *better* than the Windows alternative, because people will tend to go with Windows if they think "Alternative + Learning Curve = Effort.

Chrome isn't "just Linux," unless Linux does its device interfacing and configuring in very different ways that I'm aware of. I don't claim to be an expert--maybe it does. You make it sound, however, as if Chrome was basically a GUI with some HTML5 support baked in. Were that the case, I think the project would've been a little simpler. :P

+ -

Joel H:
You make it sound, however, as if Chrome was basically a GUI with some HTML5 support baked in.

That was my understanding of it.

+ -

Joel H:

...Chrome isn't "just Linux," unless Linux does its device interfacing and configuring in very different ways that I'm aware of. I don't claim to be an expert--maybe it does. You make it sound, however, as if Chrome was basically a GUI with some HTML5 support baked in....

A quick peek under the covers and through the image shows that it's the Linux 2.6.30 kernel, and uses the standard device drivers.

Take a look at Chrome's USB device drivers....

evil@chrome:~/c3$ ll lib/modules/2.6.30-chromeos-intel-menlow/kernel/drivers/usb/host
total 100
-rw-r--r-- 1 root root 18324 2009-11-19 18:06 isp116x-hcd.ko
-rw-r--r-- 1 root root 19407 2009-11-19 18:06 isp1760.ko
-rw-r--r-- 1 root root 24201 2009-11-19 18:06 oxu210hp-hcd.ko
-rw-r--r-- 1 root root 21567 2009-11-19 18:06 r8a66597-hcd.ko
-rw-r--r-- 1 root root 11732 2009-11-19 18:06 sl811-hcd.ko

And here's what you would see on a slightly newer Ubuntu (Lucid Lynx) box:

evil@mars:~$ ll /lib/modules/2.6.32-5-generic/kernel/drivers/usb/host/
total 400
-rw-r--r-- 1 root root 24640 2009-11-23 08:25 hwa-hc.ko
-rw-r--r-- 1 root root 34336 2009-11-23 08:25 isp116x-hcd.ko
-rw-r--r-- 1 root root 44512 2009-11-23 08:25 isp1362-hcd.ko
-rw-r--r-- 1 root root 35088 2009-11-23 08:25 isp1760.ko
-rw-r--r-- 1 root root 45328 2009-11-23 08:25 oxu210hp-hcd.ko
-rw-r--r-- 1 root root 37240 2009-11-23 08:25 r8a66597-hcd.ko
-rw-r--r-- 1 root root  9960 2009-11-23 08:25 sl811_cs.ko
-rw-r--r-- 1 root root 22672 2009-11-23 08:25 sl811-hcd.ko
-rw-r--r-- 1 root root 64280 2009-11-23 08:25 u132-hcd.ko
drwxr-xr-x 2 root root  4096 2009-11-26 10:30 whci
-rw-r--r-- 1 root root 65472 2009-11-23 08:25 xhci.ko

And Chrome puts it's pants on via init just like any other Linux:

evil@chrome:~/c3$ cat etc/init.d/
...[shell path deleted]...
export USER=chronos
export LOGNAME=${USER}
export SHELL=/bin/bash
export HOME=/home/${USER}
export DISPLAY=:0.0
export PATH=/bin:/usr/bin:/usr/local/bin:/usr/bin/X11
export XAUTHORITY=${HOME}/.Xauthority

/usr/bin/xauth -q -f $HOME/.Xauthority add :0 . $1
exec /usr/bin/ck-launch-session /etc/X11/chromeos-xsession

At the end of the day, it's another Linux distro with an minimalized/streamlined startup process and custom desktop.  Nothing related to device interaction has been fundamentally changed.

The Linux startup process has never been written in stone anyway, and can vary quite a bit from distro to distro.  With projects like Upstart, it's just getting better all the time.  The Ubuntu guys are targeting 10 seconds on a moderate system, and mine already boots in 13 - so I believe they'll get there.

+ -

>> One final bit of irony. Over ten years ago, Microsoft was sued for bundling a browser with Windows 98.  Does it amuse anyone else that Google is bundling an operating system along with their browser?


Microsoft was sued for abusing their OS monopoly.  They used the OS as leverage to bankrupt the browser companies, after the NetScape guys refused to sell the browser to MS.  And they weren't just sued... they were losing the case so badly that they paid $750M to settle.

It probably didn't help them that their internal memos went like this:

"We are going to cut off their air supply. Everything they're selling, we're going to give away for free." -Paul Maritz, former Microsoft Vice President, referring to Netscape

"It seems clear that it will be very hard to increase browser market share on the merits of IE 4 alone. It will be more important to leverage the OS asset to make people use IE instead of Navigator."
-Microsoft Manager Christian Wildfeuer (from an internal memo dated 02-24-97)

Google does not have a monopoly in the browser or OS market, therefore antitrust rules do not apply.

The sad thing about it is that Microsoft's strategy worked:  They pretty much killed Netscape and then... let IE stagnate.  This was harmful to us as consumers.  There was no browser innovation until Opera and Firefox came along.  FireFox, being free and open source, couldn't be run out of business - so MS had to put together a new IE team to make a new browser.

You might ask "Why did MS run NetScape out of the browser market... and why do they continue to compete with Firefox.... if they didn't really want to be in the browser building business?".  The answer is pretty simple and the same for all of Microsoft's actions:  Control of the API's that are used by developers.  If people started writing to the open web APIs, then there would be no dependence on the underlying OS - this is why MS encouraged ActiveX and hates Java - ActiveX forces you to write to the Windows APIs, and makes the apps specific to that platform.

Login or Register to Comment
Post a Comment
Username:   Password: