[slackline]

But now Google are to release their own operating system.

Interesting points...

• They're using the Linux kernel (this is the bit that gets hardware to talk to each other, under most current GNU/Linux OS's its the GNU bit that provides the key software to get things running, and then there are tons of other things like Xorg and various desktops and apps on top of that).
• Following on from above it seems they're going to ditch the traditional Xorg window system and write their own.
• Its going to support the ARM chips, these are used in lots of embedded systems like NAS's and routers, and they are slowly gaining market share and challenging Intel's Atom processors for netboooks
• Its different from Android, which is aimed specifically at mobile devices and netbooks

Could be quite interesting to keep an eye on this (I will be  ).

[lagerstarfish]

We hear a lot from our users and their message is clear — computers need to get better. People want to get to their email instantly, without wasting time waiting for their computers to boot and browsers to start up. They want their computers to always run as fast as when they first bought them. They want their data to be accessible to them wherever they are and not have to worry about losing their computer or forgetting to back up files. Even more importantly, they don't want to spend hours configuring their computers to work with every new piece of hardware, or have to worry about constant software updates. And any time our users have a better computing experience, Google benefits as well by having happier users who are more likely to spend time on the Internet.

Makes sense 

[slackline]

We hear a lot from our users and their message is clear — computers need to get better. People want to get to their email instantly, without wasting time waiting for their computers to boot and browsers to start up. They want their computers to always run as fast as when they first bought them. They want their data to be accessible to them wherever they are and not have to worry about losing their computer or forgetting to back up files. Even more importantly, they don't want to spend hours configuring their computers to work with every new piece of hardware, or have to worry about constant software updates. And any time our users have a better computing experience, Google benefits as well by having happier users who are more likely to spend time on the Internet.

Makes sense 

Where's the fun in that   ???  

[Palomides]

Cue everyone at gOS jumping up and down going "shitshitshit".

I wonder if their "Linux kernel" is going to really be linux, or some sort of Ginux/Goonux/gUnix in-house thing.

On the one hand it'd be great to have a linux-based OS with the weight of Google behind it, but if they googleify the kernel too much then there'll be major headaches trying to recompile anything vaguely useful to work with their windowing system. The not-quite-Java Android SDK apparently encourages a sort of benign lock-in, with porting from Java to Android being easier than Android to Java.

And on the third hand, it's all good when Microsoft get taken on directly (I bet Steve Balmers head exploded when he heard this°

[slackline]

The term Linux is commonly misused to refer to the plethora of OS's that exist out there.

The correct term is GNU/Linux, since the Linux part refers to C code that gets hardware to talk to each other (viz. drivers etc.).  The GNU part refers to operating system, things like the shell (bash, ash, zsh etc.), grep, .

See Linux and GNU and Why GNU/Linux.

So the kernel itself will be Linux, Google are just going to use that, because it already supports tons of hardware that's already out there being used, and will continue to be added to with support for new devices that come out.  They'll no doubt have some people working on drivers for new devices (or just throw some money to the GNU/FSF).

Unless of course what you're getting at is that they're going to fork the whole Linux kernel use it as a base point and add to it, which to my mind seems pretty pointless because the GPL (although I think currently the LInux kernel uses v2 as opposed to the linked v3) would mean that any modifications would have to have the source available to all on request, and they'd then get absorbed back into the main kernel.  


What Google appear to be doing is ditching the GNU part of the GNU/Linux operating system and googlifying that to their own desire.  They don't need to waste time/money/effort getting hardware working with each other as it already does (although it can no doubt be encouraged to do so faster).  Google want to integrate the desktop seamlessly with the net services that they provide so are focusing on getting it all working nicely with online backup quick booting and network connection etc.

The recompiling thing isn't really an issue either, since there will be a C compiler (whether they opt to use Intel or the GCC) and the kernel will just be built against a certain version of the compiler and the software will be too.  If they go for a binary distribution (highly likely since their stated aims are ease of use and users not having to update drivers or worry about that stuff) then they'll sort out compatible libraries at their end (I have systems with kernels compiled against one version of GCC say 4.2.2 and software running on that system built against another version say 4.3.3, its not a problem (although when you get into differences between major versions you start getting headaches, but as long as the libraries for both exist its not too much hassle, especially if you compile things statically as opposed to dynamically linking them against the libraries).

[Palomides]

Hm, I'm still trying to get my head round GNU/Linux architecture. As I (thought) I understood it, the kernel sits right at the lowest level and handles hw stuff, but for user interface there's the X Server (providing basic graphics functions), with a window manager on top.

Currently, taking one version of *buntu as an example, I can choose between xubuntu, kubuntu and ubuntu, which contain the same linux kernel, the same X server and three different desktop interfaces (plus their window managers): Xfce(with xfwm), KDE (with Kwin) and GNOME (with Metacity).

I thought (this is the bit where I'm kinda guessing) that when you develop a GNU/Linux application, you have to choose which GUI toolkit to use - and therefore which desktop interface the application can run on??

If that's correct, then doesn't that mean if Google release a Linux-based OS with a brand new windowing system, then existing Linux applications will need to be re-written and recompiled with whatever GUI toolkit Google supply, to be able to run on the Google OS???

[slackline]

No not really at all.

To start with with the *buntu family refers to the desktop software, each of which has its own window manager Kbuntu == KDE Xubuntu == Xfce Ubuntu == GNOME.  KDE/GNOME/Xfce are all desktop environments, they each require a window manager and Metacity is the default one with GNOME, but all will work with say Compiz as the WM if you want fancy graphical effects, or fvwm if you want to use the window manager from Xfce to manage windows in your GNOME session!

Anyway, theres a layer missing out.

Most applications have CLI's (Command Line Interfaces), a GUI is then provided in either Qt (for KDE) or GTK (for Gnome or Xfce).

Take for example transmission a torrent client.  It runs as a daemon, but you can have a stand alone GUI to it in either GTK or Qt, or you can forget completely about the GUI, as the daemon has a built in http interface so you can access it via a WebUI.

Another example is pyBibliographer which is a piece of software for managing scientific bibliographers.  It can be run at a terminal no problems, but because its written in Python and there is an API between Python and GTK a GUI is provided that allows you to use a nice GUI.

Personally I abhor KDE its like a cartoon for kids, but there are some apps (e.g. the excellent CD/DVD burning tool k3b) that I use that only have GUI's via Qt so I have Qt libraries installed that mean I can run these apps under Xfce4 or even fluxbox (which I use on my laptop 'cause its really light-weight).

The beauty of all this is that whether you are using KDE/GNOME/Xfce4/fluxbox/Openbox and their attendant window managers and toolkits (Qt/GTK) they are both of Xorg.  If you've ever done a real basic install of GNU/Linux then after you've installed Xorg you're generally advised to test it out with startx, you then get two (or three) terminals, a clock and some other things.

I don't think I've explained that very well though (I've confused myself just writing it, its one of the minor nuances, and major advantages, of GNU/Linux that everything is so modularised and fragmented as opposed to being monolithic), and having just re-read the Google blog they say...

The software architecture is simple — Google Chrome running within a new windowing system on top of a Linux kernel. For application developers, the web is the platform. All web-based applications will automatically work and new applications can be written using your favorite web technologies.

So I was perhaps wrong to say that they're going to ditch Xorg, what they're going to do is write their own version of Metacity (the window manager for GNOME)/FVWM (the window manager for XFCE4)/fluxbox/compiz (full list of window managers available under gentoo).

All a window manager does is control the placement and apperance of windows this is probably worth reading instead of me confusing myself further!  But I very much doubt there will be a problem.
Further you won't have to write applications for the GoogleOS, they'll all be web-based so its nothing to worry about.

[Palomides]

Brilliant - that makes things a lot clearer, thank you. I hadn't understood the independance of all the components & libraries, and the way they can co-exist inside the same installation.

But all that may be just an interesting but irrelevant tangent. doesn't this:

The software architecture is simple — Google Chrome running within a new windowing system on top of a Linux kernel. For application developers, the web is the platform. All web-based applications will automatically work and new applications can be written using your favorite web technologies.

Sound a bit like "you can run Google Chrome and nothing else"?

[slackline]

Brilliant - that makes things a lot clearer, thank you. I hadn't understood the independance of all the components & libraries, and the way they can co-exist inside the same installation.

  I made sense?  I thought I was just muddying things further 

But all that may be just an interesting but irrelevant tangent. doesn't this:

The software architecture is simple — Google Chrome running within a new windowing system on top of a Linux kernel. For application developers, the web is the platform. All web-based applications will automatically work and new applications can be written using your favorite web technologies.

Sound a bit like "you can run Google Chrome and nothing else"?

I'm not so sure it does.

The way I see it, Google are trying to revert to the way computers used to run, a company would buy a main frame which packed all the data storage and processing power in one box hidden in a room somewhere in the depths of a building, and people then logged on via a client terminal and shared the resources.

As Moore's law progressed things got smaller and cheaper and IBM/Apple came along and produced desktop machines.  M$-Windows provided an OS that "simplified" (in its very loosest term) interacting with the computer, and over the last couple of decades its all been about packing more punch on your desktop (faster CPU, more RAM, more GB of HD).

But somewhere along this line came Tim Berners-Lee with his wonderful idea of the WWW sitting on top of the internet that already existsed.  This meant that content rich information could be shared between people.  A decade or so later along come Google, who now dominant not just the search market but are starting to revert to the old model of computing (glossing over lots of steps here obviously, but the point is that we now share content over the network that is the Internet).

Look at Google Docs, you now write your report online and can share it with others, access it from anywhere others who are permitted to do so can edit it too.  Its reverting back to having everything centralised like in the old days, and thats where Google Chrome OS comes in, its going to be geared towards not needing to install applications locally, it will all run on Googles server farms, what the OS will focus on is getting the computer up and running and connected as quickly as possible, so you're not waiting around whilst it boots, or waiting for OpenOffice to start up or whatever.

I don't think this will mean that you won't be able to run say an mp3 player locally under Google Chrome OS, although they may take the approach of providing say online storage for your music which you can stream anywhere (kind of like lastfm do already), but if all they're doing is writing their own window manager that won't be a problem.  WebApps will (I'd imagine) work under say IE running on XP, but perhaps not as fast as they would under Google's Chrome OS (or browser) as there won't be as tighter integration (at least until M$ get up to speed and utilise the API to its full extent).

All in all though its good that this will be a serious challenge to M$'s monopoly.

[Palomides]

Some more information here

http://www.computerworld.com/s/article/9135293/Opinion_Why_Google_Chrome_OS_matters_already_on_Day_1

If this guy knows what he's talking about, then it answers a couple of things I was wondering about:

Google gears to allow switching between online/offline working (integrated in at a low level for high performance?)

Totally new desktop interface but with APIs from the Portland Project to allow existing Linux applications to run.

I guess Google's biggest challenge is going to be how well the whole thing adapts to the switch between online and offline. It's all well and good saying how this should be transparent for the user, but if I'm moving around with a laptop, then I need control over network connectivity, especially wireless. I need to be able to turn off network connections in aircraft, shut down wifi hardware to save battery power when hanging around, and I definitely don't want to be hooking up to any old wireless access point and throwing all sorts of google account information around.

Actually I think google logons use SSL by default, but that's still vunerable to man-in-the-middle attacks - no "Public Free Wifi" for me thanks. And yes, I do have gmail set to use SSL for the whole session, but it's not the default set by Google, and I don't think it's supported for Docs yet?

[slackline]

Good article.

Agree that it will have to be intelligent enough to allow moving around, choice of connection etc.

Actually I think google logons use SSL by default, but that's still vunerable to man-in-the-middle attacks - no "Public Free Wifi" for me thanks. And yes, I do have gmail set to use SSL for the whole session, but it's not the default set by Google, and I don't think it's supported for Docs yet?

Thats something I only realised a month or so ago (that SSL for the whole session wasn't default).  For anyone using gmail accounts etc its well worth changing this setting (Just go into settings and under "Account" at the bottom is the option to always use https)

[slackline]

Google Chrome blog

Seems there are lots of industry players falling over themselves to get a slice of the cake.


Interesting analysis

Good point about people potentially having to start with a "clean slate" if/when they switch from M$.

[Drew]

I need a smiley for either my head exploding, or me blowing my brains out. Either will suffice!

[butters]

Thats something I only realised a month or so ago (that SSL for the whole session wasn't default).  For anyone using gmail accounts etc its well worth changing this setting (Just go into settings and under "Account" at the bottom is the option to always use https)

It is under Settings --> General on my Gmail account - it is at the bottom though as you say.

[slackline]

Thats something I only realised a month or so ago (that SSL for the whole session wasn't default).  For anyone using gmail accounts etc its well worth changing this setting (Just go into settings and under "Account" at the bottom is the option to always use https)

It is under Settings --> General on my Gmail account - it is at the bottom though as you say.

 

[dobbin]

Android now avail as a live CD : http://www.theinquirer.net/inquirer/news/1433412/the-android-os-comes-pc

[lagerstarfish]

Am I right in thinking that it might be a good idea (in the f u t u r e) to partition my hard drive and have Go Ogle Chrome OS on one bit for all my 'tinterweb needs and have some sort of windows shite on another part for all the mainsteam programs I might want to use?

[slackline]

Am I right in thinking that it might be a good idea (in the f u t u r e) to partition my hard drive and have Go Ogle Chrome OS on one bit for all my 'tinterweb needs and have some sort of windows shite on another part for all the mainsteam programs I might want to use?

Quite possibly, although you'd likely end up getting pissed off with booting between the two OS's and end up running Chrome OS (or some other Free Open Source Software OS  ) under an emulator such as VMware or VirtualBox within your OS of choice which runs the mainstream programs you want to use.  That said you wouldn't benefit from the proposed fast booting aspects of the Chrome OS if you were switching between it and your OS of choice by dual-booting or running Chrome OS under VMware/VirtualBox, which is really missing the point of one of the purported aims(/benefits).

[lagerstarfish]

I'd like quick booting sessions for webshite and would be quite happy to wait for windows to boot up (with no internet connection) for more involved stuff especially if that means fewer virus thingys

[slackline]

That's cool then