Daring Fireball: Maybe Instead of Two Cars, You Just Need a Car and a Bicycle
daringfireball.net
daringfireball.net
Apparently the world consists only of people with two computers. Nobody has a single computer that is overseeving their needs and it is inconcievable that there are people who simply haven't bought one yet, perhaps because they are overwhelmed by their complexity.
The easiest way to build a product that kicks ass is to start with someone else's great idea (camcorders, for instance), and take stuff away.
I travel a lot, and I do a lot of programming when I travel. So, I bought an Acer Aspire A150, I got a spare 9 cell battery and I increased the RAM to 2 GB. And, it runs Windows 7 Ultimate like a charm. I have enough battery life to last me a 14 hours of coding movie watching and non-stop music listening. The whole thing is the size of a hard-cover book.
I use Rhapsody for music, which keeps my music libraries synced and Bitbucket for my code. I use Windows mesh to keep other files synced to the "cloud". I can switch seamlessly from my 3 monitor dev box at home to my netbook by doing an "hg pull, hg update"
How does using a "cloud centric" OS help me? Is it cheaper? I got Win 7 Ultimate for free through Bizspark and I have $350 invested in hardware. I don't really see the advantage of hamstringing my OS. Hardware is cheap.
I understand the theoretical advantage to the third world of using less hardware, but when I was in Mexico this spring, netbooks were everywhere. They weren't running Ubuntu, either, they were running pirated copies of Windows.
I'm not trying to troll, but aside from trying to compete with MSFT, how is a cloud centric OS any better than a typical OS?
1. Web apps are a plus for a secondary device, because you won't need to copy files back and forth, or install sync/VCS software (a hassle for non-expert users). HTML5/Gears storage remove some of the disadvantages by letting you continue work offline. (On the other hand, services like Dropbox have made it just about easy enough for non-expert users to do this with traditional "local" files.)
2. No user-installed native code means certain security concerns could be greatly reduced. All installed code is signed, verified on boot, mounted on a read-only partition, and can be restored to a pristine state by the firmware. Web apps are subject to browser same-origin policies and to Chrome's process isolation. Downloaded native code can run only in the NaCl sandbox. A broken app can't trash all your stuff, because data from each web site is accessible only to that site. (On the other hand, web apps are almost universally vulnerable to XSS and XSRF and to server-side exploits. NaCL has been broken before and will be broken again. The web security model is a start, but it has flaws both in design and in practice.)
3. The interface is simplified in a few key ways. Instead of windows that represent documents within apps and often have their own separate tabs or pages, everything the user interacts with is a tab within a browser. Not a big deal for savvy computer users, but for novices this is one less concept to learn. Everything is a web page; they don't need to be "installed"; they all appear in the tab strip and have addresses that can be bookmarked and emailed. (On the other hand, web apps share only the most basic UI conventions; they are often radically inconsistent from one to the next.)
4. The Chrome team's obsession with speed may prove to be one of those subtle, hard-to-explain benefits. (Like trying to explain why Apple's obsession with detail matters, to someone who hasn't used their products.) I never had a problem with Firefox until I used Chrome for a few days; now any other browser seems painfully unresponsive. Same thing with git compared to darcs/bzr/svn. When more operations move from "fast" to "instantaneous" it changes the feel of the product, and that affects usage patterns.
5. Compared to Windows, Mac OS X, or normal GNU/Linux distros, Chrome OS has less overall code, is conceptually simpler, and has few legacy/compatibility constraints. If it's managed right, this could mean that it will be faster to evolve and optimize, easier to secure, and ideal for experimenting with new technologies and user interfaces. It will also help motivate next-generation web applications and web standards.
Think about people in countries where it's not common to have an iPod that you want to sync 50GB of music to. Or very casual users.
But eventually there will be Web apps as good as the local programs for doing what you want: BitBucket, for example, could launch an online editor that integrates with them.
Think how much people use the internet, email and chat. I use my netbook almost exclusively for those things - and I dont think I am alone.
I really do think you are missing the point mainly due to being a power user / hacker. A lot of my friends who are casual computer users a) have netbooks and b) are a perfect target market for this.
Some are already asking me about Chrome OS with interest.
Try to think of it as if your a joe user with little technical knowledge.
Limits to some of us are freedom to others. It all depends on perspective.
One of the things that Google appears to be focusing on most intently with ChromeOS is syncing the state of your workspace environment. If they can deliver on that, it'd make it worthwhile to own devices in dissimilar form-factors: like a 1280x800 laptop, a 1024x600 Touchbook, a 800x480 Pandora, and a 320x480 iPhone 3Gs (the latter three all ARM-based!). I'd even use it on a 1920x1200 desktop if they implement a decent tiling idiom.
All I really need is a very nice web browser and some clever NaCL ports.
For most of what people want to do, a netbook has enough or almost enough horsepower. Heck, even an iPhone has enough power for most of it. The only thing you need a full-blown desktop for is for heavy duty Photoshop, graphics software, gaming, or something demanding like that.
So, note how the form factors are disconnected from capabilities:
- iPhone - capable of word processing, but lacks a keyboard to do a lot of
it comfortably
- All-in-one desktop machine - often capable of demanding gaming, but
marketed as the email/web machine for your grandmother.
I see computing coming in two major forms in the future: highly portable, or home/office integrated.Much of our personal data will be on devices sized like the iPhone or Blackberry. They will be small enough to go everywhere with us. What doesn't fit there will live in game-console like machines in our homes, or on office workstations/high-end laptops, and will move around using the network. As networks get better, much of that will live in the "cloud."
Given better form factors (dockable screens and keyboards and other, newer input devices) the "go-everywhere" device will subsume more and more of our computing tasks. There will no longer be netbooks. There will be "things that allow us to word process/schedule/surf more comfortably" by providing better form factors for interface, but the data will live in our pocket and be synced with the cloud. There will be "other things with the compute/graphics/storage horsepower for specialized tasks." These will be the "hubs, consoles, workstations." These will be the only stationary things, because they will be the only things that have to be stationary.
Everything else will be interface. (The OQO will live at last!)
The Chrome OS webcast made repeated references to data cached locally. I'm not sure but it sounded like it might even cover more than just apps specifically written for Gears/HTML5.
Smartphones are in a position to be your second computer in the next few years. Personally I prefer the smartphone but I can see some people wanting a web only computer.