[The Final Dakar]

starman seems be slowing morphing into this forums Dvorak. (Not that his points are without merit)

[TETENAL]

Originally Posted by starman 

Google's creating an unsecure OS, which relies on the 'net, and doesn't run anything except what's on the 'net.

Surfin’ Safari - Blog Archive � WebKit Does HTML5 Client-side Database Storage

[Jawbone54]

There is zilch about Chrome OS that excites me. I'll check it out if it's still around in 3 more years.

[The Final Dakar]

Best I can tell it's a niche OS for a niche market. But given what google is trying to pull here, that seems like the smartest place to start.

[nonhuman]

A simple example of the awesome **** HTML 5 will enable us to do: A Platform Game Using HTML5 Canvas � PurpleFloyd’s Blog

[TheoCryst]

Originally Posted by nonhuman 

A simple example of the awesome **** HTML 5 will enable us to do: A Platform Game Using HTML5 Canvas � PurpleFloyd’s Blog

Okay, that's pretty damn impressive. Maybe Google is on to something after all. When I'm honest with myself, I can't think of a single thing that you would want to do with a netbook that you couldn't do on one of these.

Now I won't be replacing my MBP any time soon with one of these, but that doesn't mean I don't see their potential. This is Google's first push towards their ideal world: a world of thin clients tied to their online services. There will still be plenty of room for rich clients in the foreseeable future, but I expect to see those devices slowly marginalized over the next couple of decades. After all, did anyone really predict that a web browser could replace the entire OS fifteen years ago? Of course not, the concept would have been laughable. But here we are, having an actual honest-to-God debate about the possibility. The times, they are a-changin'.

[adamfishercox]

From what I see, the Chrome OS is awesome in that it's based entirely on the web. When compared to other existing OSes, there's not much great about it from a consumer point of view, except for that it is created with netbooks in mind.

[starman]

Originally Posted by adamfishercox 

From what I see, the Chrome OS is awesome in that it's based entirely on the web. When compared to other existing OSes, there's not much great about it from a consumer point of view, except for that it is created with netbooks in mind.

So then what makes it any different than a stripped down version of Linux that just autolaunches Firefox?

[adamfishercox]

I guess the fact that it's built for a specific type of computer in mind rather than just being a general OS with a browser.

The main point of my post is to say I see no worth in it, just as I see no worth in netbooks. Doesn't mean it wouldn't work great for some.

[starman]

I see a lot of worth in netbooks, but not Chromium.

I played with it for about 15 minutes. It's just a browser (duh). THAT'S IT. It does NOTHING more.

Google thinks this appeals to people? Just get Linux for God's sake, at least it can do more than browse the 'net.

[The Godfather]

Because it works without a bootloader, it will be impossible to have a dual/triple/quad boot netbook if it comes with Chrome's special firmware. No more BIOS or GRUB.

[lpkmckenna]

I think Chrome OS is a great concept. Why bother with a mail app when there's GMail and Hotmail and Yahoo mail, or why bother with an office suite when there's Google Apps or Windows Live Office? So much of the user-facing functionality of an OS can be done in the browser, so there's no reason not to try it.

How many of the basic apps of Mac OS X couldn't be implemented as web apps? TextEdit, Preview, QuickTime, ImageCapture, PhotoBooth, even iTunes could be replaced. (In fact, the iTunes clone SongBird is a media player build out of Gecko.) Hell, Dashboard is really nothing but a foundation for web apps.

Frankly, most Linux distributions are too heavy for netbooks. Performance in OpenOffice isn't great on a regular computer, never mind a netbook.

[CharlesS]

Originally Posted by starman 

Google thinks this appeals to people? Just get Linux for God's sake, at least it can do more than browse the 'net.

And you think Linux appeals to people?

Actually, from the sound of it, Chrome OS is Linux, so you probably would be able to figure out a way to install an Xterm, Gnome, etc. if you so wanted.

[lpkmckenna]

Originally Posted by The Godfather 

On the other hand, if this was targeted at replacing Android, rather than revolutionizing the laptop, it would make sense.

Chrome OS isn't about revolutionizing the laptop, Chrome OS is designed only for netbooks. Android is designed for phones. Can you imagine using the tab bar on a phone?

MS and Apple will surely invent something that will blow Chrome out of the water in 2010.

Sorry, but no. Apple is incapable of making free services on the web (except their movie trailer site and dashboard/web apps pages), and Microsoft currently makes the slowest, least capable browser on the market.

[lpkmckenna]

Originally Posted by CharlesS 

Actually, from the sound of it, Chrome OS is Linux, so you probably would be able to figure out a way to install an Xterm, Gnome, etc. if you so wanted.

Probably not. Chrome OS isn't a complete distribution, so no x11. You would pretty much have to install an entirely new Linux distribution on top. It would be easier to code a terminal web app.

[CharlesS]

We used to have to install a complete X11 distribution on OS X too in the old days, and we managed.

[besson3c]

Originally Posted by lpkmckenna 

I think Chrome OS is a great concept. Why bother with a mail app when there's GMail and Hotmail and Yahoo mail, or why bother with an office suite when there's Google Apps or Windows Live Office? So much of the user-facing functionality of an OS can be done in the browser, so there's no reason not to try it.

How many of the basic apps of Mac OS X couldn't be implemented as web apps? TextEdit, Preview, QuickTime, ImageCapture, PhotoBooth, even iTunes could be replaced. (In fact, the iTunes clone SongBird is a media player build out of Gecko.) Hell, Dashboard is really nothing but a foundation for web apps.

Frankly, most Linux distributions are too heavy for netbooks. Performance in OpenOffice isn't great on a regular computer, never mind a netbook.

A significant ramification out of all of this I can foresee is the concept of legality and data ownership. This will all be fine for those who do not need to store confidential and/or private business information, but the problem is there are so many people who never stop to think that they are jeopardizing themselves by storing private company data they do not own in their GMail account.

[besson3c]

There is another ramification of netbooks like this thriving. We've talked about this a little in the past, but it's the notion of some users refusing to use apps that are not true "Mac" apps or apps that don't abide by the principles of the Mac OS interface guidelines. If platform agnostic web apps really thrive, what incentive is there for developers to conform these apps to a set of desktop app guidelines for OS X? For starters, if they did this for Windows too there would be more versions of their app to support, but more importantly it seems like there is already widespread understanding that web apps are their own platform with their own interface and their own interface guidelines.

Look at all of the thousands of users that accept the Facebook interface, or the MySpace interface. Perhaps it makes sense to think about this in the reverse, and that is to get Apple to design their apps to leverage the familiarity people have with apps like Facebook?

Things are definitely changing, and it's hard to see how the OS X (or Windows) interface guidelines really carry much importance in the future.

[slugslugslug]

Right, besson3c is finally hitting the one big problem I see with Chrome/Chromium OS. It’s great in a sense that people can make web apps that run on Mac & Windows & Linux & netbooks. But because they can just throw together any UI and skin, it makes the whole idea less appealing to me. I can’t imagine using a webmail interface as my primary e-mail manager or Google Reader as my primary RSS reader because it’s such a pain in the ass to figure out where everything is. When people make an app targeted to my OS, I know or can guess ahead of time what several keyboard shortcuts are, and I can find others I’m looking for really easily by navigating menus. And visual consistency is useful too, even if not essential (Yes, I can recognize a close-window/tab/whatever widget in a web app, since it’ll be something with an “x” in it, but I prefer when those kinds of things are more similar).

Anyway, I’m intrigued by this whole idea, but I doubt I’d be convinced to get a Chrome netbook, even as a secondary computer.

[besson3c]

Originally Posted by slugslugslug 

Right, besson3c is finally hitting the one big problem I see with Chrome/Chromium OS. It’s great in a sense that people can make web apps that run on Mac & Windows & Linux & netbooks. But because they can just throw together any UI and skin, it makes the whole idea less appealing to me. I can’t imagine using a webmail interface as my primary e-mail manager or Google Reader as my primary RSS reader because it’s such a pain in the ass to figure out where everything is. When people make an app targeted to my OS, I know or can guess ahead of time what several keyboard shortcuts are, and I can find others I’m looking for really easily by navigating menus. And visual consistency is useful too, even if not essential (Yes, I can recognize a close-window/tab/whatever widget in a web app, since it’ll be something with an “x” in it, but I prefer when those kinds of things are more similar).

Anyway, I’m intrigued by this whole idea, but I doubt I’d be convinced to get a Chrome netbook, even as a secondary computer.

There are many that feel the same way as you do, but as every day passes and I come to learn more about just how big apps like Facebook have become, I feel that more and more this will be a losing battle.

Any UI person will tell you that a big part of communicating concepts to users is leveraging what is familiar to them. There are more people that will be familiar with Facebook than with OS X itself. Hell, there are probably more people that are familiar with Facebook now then they are MS Office.

Because there isn't one dominant social network the UI of the future will probably involve taking common ground from these apps. Perhaps the "Mac way" will influence some of these apps, but really, Apple is in no position to be controlling the future of UI at this point, unless the iPhone ends up competly dominating the entire phone market and becomes an interface even more familiar to people than these social networks.

My point is that these sorts of minority groups will have to make computing concessions - that is, accepting things that aren't absolutely perfect for them. There are only a handful of companies that will devote resources to catering towards minority groups. There may be enough to create a very health eco-system just as there is in OS X today, but the problem is that there will be so many more minority groups to cater towards with things changing as they are and the desktop becoming less and less relevant:

- all of the major smart phone interfaces
- all of the major social networks
- business software interfaces - whether this is cloud or desktop
- other desktop software interfaces

[shifuimam]

Originally Posted by besson3c 

Look at all of the thousands of users that accept the Facebook interface, or the MySpace interface. Perhaps it makes sense to think about this in the reverse, and that is to get Apple to design their apps to leverage the familiarity people have with apps like Facebook?

Things are definitely changing, and it's hard to see how the OS X (or Windows) interface guidelines really carry much importance in the future.

Apple maybe, but not Microsoft. The big money from Windows is in OEMs, and the big money for OEMs is in corporate contracts.

Sure, an end user such as yourself may be okay with keeping all your private files on someone else's server out in the "cloud", but a business with serious intellectual property concerns isn't going to be quite so interested in the concept of their data being on a server whose security is out of their control.

Not only that, but what happens when your Internet connection dies? What happens when the line from your office to your datacenter goes out, taking down the Internet and LAN connections with it? People can't even rely on their own internal servers being constantly available, let alone some server in la-la land on the "cloud".

My boyfriend recently started working as an IBM contractor on a project for the TSA. They gave him this little thin client laptop that's completely useless without being connected to what is essentially the "cloud" - the internal mainframe. They also are giving him an actual laptop so that he can do development work at home and while traveling without needing a reliable, high-speed connection to the Internet to do so. While there are benefits to the thin client (and the desktop he has at work that runs Windows but also connects to a Sun mainframe that allows him to access highly confidential systems, data, and websites), such as the high security, it is not an end-all solution to the IT needs of a corporation.

And, while it doesn't seem like it to you (especially because it's never really been one of Apple's focuses), corporate IT is by no means a "minority group" in the world of software, technology, operating systems, and hardware.

[besson3c]

Originally Posted by shifuimam 

Apple maybe, but not Microsoft. The big money from Windows is in OEMs, and the big money for OEMs is in corporate contracts.

Sure, an end user such as yourself may be okay with keeping all your private files on someone else's server out in the "cloud", but a business with serious intellectual property concerns isn't going to be quite so interested in the concept of their data being on a server whose security is out of their control.

Not only that, but what happens when your Internet connection dies? What happens when the line from your office to your datacenter goes out, taking down the Internet and LAN connections with it? People can't even rely on their own internal servers being constantly available, let alone some server in la-la land on the "cloud".

My boyfriend recently started working as an IBM contractor on a project for the TSA. They gave him this little thin client laptop that's completely useless without being connected to what is essentially the "cloud" - the internal mainframe. They also are giving him an actual laptop so that he can do development work at home and while traveling without needing a reliable, high-speed connection to the Internet to do so. While there are benefits to the thin client (and the desktop he has at work that runs Windows but also connects to a Sun mainframe that allows him to access highly confidential systems, data, and websites), such as the high security, it is not an end-all solution to the IT needs of a corporation.

And, while it doesn't seem like it to you (especially because it's never really been one of Apple's focuses), corporate IT is by no means a "minority group" in the world of software, technology, operating systems, and hardware.

You're absolutely right. I have a pretty good awareness of the notion that many businesses are going to be weary of putting their data up in the cloud, I'm personally weary of this in my own business, in part because people like to do things such as email me their passwords

My assumption all along is that these Google Netbooks are going to be targeted at home users, at least at first. Large businesses can buy their own Google search appliances, perhaps in the future (if not already) Google will allow large businesses to run the Google apps locally on similar appliances?

[shifuimam]

Even so, there's a risk of major loss of productivity whenever the corporate "cloud" goes down or has some technical problem. In a very large business, you may only have one or two datacenters in the entire country. If something happens to a site's connection to one of those datacenters and you're doing something like Chrome OS, suddenly none of your employees can do anything. That's a pretty major problem in a world where time is money.

Also, weary != wary.

[besson3c]

True, that dependency has its downsides...

Thanks for the wary correction, I weary appreciate it!

[Chooglin']

It's amazing how many people can be so short sighted.

ChromeOS and other open source software are the future. The propietary computer desktop model that Microsoft and Apple have profited heavily from is dying. Both companies know that open source software is rapidly advancing and producing better applications than their business model can possibly hope to produce. Just sit back and watch the implosion.

[The Godfather]

While I agree that Open Source is the future for operating systems, I don't think that cloud computing has more application than facebook-type and hulu-type internet consumption. While iPhoto can be moved to the cloud, the current and up-coming series of creative applications are better served with a local CPU and storage.

Unless the western world is becoming less creative, creative applications will continue to guarantee the dominance of high power CPUs and Terabyte HDDs.

[besson3c]

Originally Posted by The Godfather 

While I agree that Open Source is the future for operating systems, I don't think that cloud computing has more application than facebook-type and hulu-type internet consumption. While iPhoto can be moved to the cloud, the current and up-coming series of creative applications are better served with a local CPU and storage.

Unless the western world is becoming less creative, creative applications will continue to guarantee the dominance of high power CPUs and Terabyte HDDs.

Storage, no, CPU, yes... We are already at the point where we can make cloud computing far faster than your local PC. There is no real upper limit to the kinds of performance you can get from a computing cluster. The question is of economics and business: who will be paying for this cluster and how, but these apps are definitely not best served with a local CPU. Local video card, sure (bandwidth is a bottleneck here), but not CPU.

As far as storage goes, right now the way I see it there is still a great divide between enterprise and consumer storage. If you purchase a VPS from Linode or Slicehost you'll get a generous amount of bandwidth, CPU, and memory for a relatively modest amount of storage. These sorts of clusters are typically setup with large SANs and other disk arrays that utilize expensive SCSI/fiber channel disks rather than cheap SATA. You cannot get a 1 TB enterprise class disk for less than $100 on NewEgg.

The problem here is that some sorts of enterprise usage by its nature requires very fast I/O - email and heavy database usage, for example, whereas some usage is just long term storage. It does not make economic sense to archive email that will just sit there and never get touched onto these crazy fast disks when you simply need to have it stored and you don't really care about how fast it is accessed when it is.

We have cheaper storage based solutions such as ZFS and iSCSI sorts of things emerging, but what would be really cool is if we could write software that will automatically offload files that are rarely accessed and can safely be archived onto these slower disk arrays which can be filled with super cheap disks.

The difference in cost between a fiber channel drive and a SATA drive is immense (e.g. Seagate Cheetah 15K.6 450 GB Fibre Channel Internal Hard Drive ( ST3450856FC ) - Hard Drive - Seagate Storage). An alternative to these expensive drives might be creating a wider stripe of cheap disks, but this adds costs to your facilities, power consumption, and maintenance costs.

[The Godfather]

Storage yes, CPU no?

I can't imagine what application would be better served with a supercomputer on the cloud, rather than my 3 year old Core Duo. And I will be happy to replace it with my own dollars rather than to depend on the "economics and business" of Google and Yahoo.

Unless Pixar starts offering a "make your own CGI movie" online service, there's no possible killer app for cloud CPUs.

[besson3c]

Originally Posted by The Godfather 

Storage yes, CPU no?

SOrry, I meant that the other way around. I've corrected this.

I can't imagine what application would be better served with a supercomputer on the cloud, rather than my 3 year old Core Duo. And I will be happy to replace it with my own dollars rather than to depend on the "economics and business" of Google and Yahoo.

Unless Pixar starts offering a "make your own CGI movie" online service, there's no possible killer app for cloud CPUs.

You are mixing business with technicalities, I think. I was only addressing the technicalities in what you wrote:

the current and up-coming series of creative applications are better served with a local CPU and storage.

Technically speaking, you can get far more performance out of non-local CPUs.

[The Godfather]

Yeah. I mixed up the commercial and technical argument here. However I also argue that technically, local CPUs will always trump cloud CPUs.

While there's no upper limit to the performance of a computer cluster, this computer cluster (or clusters, all the same) is shared among all your clients. Like cable broadband, you have great performance at the beginning, but as it grabs marketshare, it becomes starved for CPU time and transfer speed. Worse yet, if any of these cloud applications is managed like AT&T's cellular network, it will take your money and not reinvest it in more network pipes and CPUs, making your experience sour and infamous.

...and I went off again with the commercial side of the argument. Sorry, man. I just think that my CPU+GPU will always have a preferential space in my heart.

[besson3c]

Originally Posted by The Godfather 

Yeah. I mixed up the commercial and technical argument here. However I also argue that technically, local CPUs will always trump cloud CPUs.

While there's no upper limit to the performance of a computer cluster, this computer cluster (or clusters, all the same) is shared among all your clients. Like cable broadband, you have great performance at the beginning, but as it grabs marketshare, it becomes starved for CPU time and transfer speed. Worse yet, if any of these cloud applications is managed like AT&T's cellular network, it will take your money and not reinvest it in more network pipes and CPUs, making your experience sour and infamous.

...and I went off again with the commercial side of the argument. Sorry, man. I just think that my CPU+GPU will always have a preferential space in my heart.

Yeah, that's not a technical argument

However, I think you're also looking at this from a mathematical relationship of more customers = slower until more resources are allocated. This is sort of true, but there is no simple curve to illustrate this relationship. An intelligent VM cluster will allocate resources on the fly, giving idle resources to hungry jobs as needed. Many adjustments are made constantly, even transferring entire VMs to different physical machines as needed.

It is very efficient in this sense. When you are not working your CPU is doing nothing but sitting in a box. Therefore, when you think about it is really pretty wasteful and expensive to have this blazing fast CPU available whenever you need it. Even for not running cloud apps and just running Desktop apps you can still get more performance and efficiency from a local VM cluster. Local meaning, say, local to an office or even a household.

That's kind of the direction I think computing is going in, starting with businesses.

[The Godfather]

Can you get more efficiency from a local VM cluster? More efficiency than the apps running natively on the host CPU? I thought that even with hardware accelerated VMs, performance is never up to par. Isn't the "efficiency" claim by VM vendor really a claim that their new product is more efficient than the previous generation of VMs. Not really a "we'll make your apps run faster" kind of promise.

[besson3c]

Originally Posted by The Godfather 

Can you get more efficiency from a local VM cluster? More efficiency than the apps running natively on the host CPU? I thought that even with hardware accelerated VMs, performance is never up to par.

Yup.

It's kind of a weird concept to wrap one's head around... We're used to the fact that a single task on a single 2.5 Ghz CPU will complete faster than a single 2.0 Ghz CPU. However, when you get into multiple cores and making far more RAM available, *if* your app is designed to make full use of these processors/cores and gobs of RAM, you can do far better providing the host CPUs are up to the job and are not resources starved. Computing seems to have moved away from the faster clock speed = the name of the game, to making more CPUs and cores available for being able to farm out more work.

[besson3c]

Also, disk speed is very important too. You'll get far more disk speed out of a VM cluster too built from fast shared storage.

[The Godfather]

I'll take your word for it, or you can provide a reference to literature that describes how these VMs work.

Back to Chrome OS, is anyone enjoying the new OS on their Mac?