Google Chrome OS. Or, how KDE and GNOME managed to shoot each other dead

Google Chrome OS. Or, how KDE and GNOME managed to shoot each other dead


A lot of people at the moment are immensely intrigued by Google Chrome OS. I won't hide that I am one of them. Google promises a much needed shift in the way small computers work. Problems like software updates, backups, installation, maintenance, viruses, have plagued the world for too long: a shift is way overdue. To me, however, the change about to happen shows us what many people have refused to believe for a long time: KDE and GNOME shot each other dead. I write this knowing full well that I am going to make a lot of people angry. This might be the first time a writer receives very angry responses from both camps -- KDE and GNOME's users might actually (finally?) join arms and fight just to show everybody how wrong I am!

So, let's go back a little bit -- not much: just a year or so. You are Google and you want to provide the operating system for the next generation of users, the ones who didn't start with Excel and Word, but with Facebook and Flickr. The obvious choice is GNU/Linux for the kernel -- Google knows it well, helps improving it, and obviously likes it. Then, the next question: what desktop environment would you feed those new users? KDE? GNOME? Both? What about programs looking different? What about the broken audio system? (Pulseaudio anybody?)

The question was a tough one. The answer was simple and painful: neither of them. Painful, because I am intimately sure (although I can't prove it) that if GNU/Linux had one set of desktop libraries, one desktop environment, one set of standard for playing audio and so on, we would have those libraries in Google Chrome OS. Google would have released a set of tools to bundle software in Chrome OS -- something without the absurd current problems of software installation in GNU/Linux.

However, two different "everything" in the GNU/Linux desktop world meant that the break from the past, in Google Chrome OS, had to be more definite and definitely more radical. Google Chrome OS, at least initially, will not allow third party software bundles. Eventually, I am 95% sure they will have to give in -- at that point, they will have to deal with the KDE/GNOME split and the result will be business as usual: messy.

Did I say that Chrome OS will allow software installation?

The first snapshots of Chrome OS are clearly showing an operating system that won't allow any third party software to be installed. Just to make sure... the root file system is mounted read only! This is a clear signal from Google (and it's a signal I like): the main file system is for the core system. There is no point in installing end-user applications there: too many problems arise if you do that. An end-user application ought to be a self contained folder full of all the rubbish the application is meant to come with.

(Of course, if there was only one desktop environment in GNU/Linux, it would be trivial to make a list of minimal libraries for each "major version" of the system. Then, something like this would be possible: drag an old icon with your 2 year old accounting software from your backup, and use it without entering dependency hell. However, GNU/Linux has two different desktop environments, so having a neat set of libraries isn't as trivial as it ought to be.)

However, only allowing web applications is going to make the transition too hard. More and more people will complain about it. A lot of users will have at least one little program they will want to be able to run on their Chrome OS laptop (I personally have one, GNUCash, And no, there is no "web app" equivalent for it...). The risk here is that Google leaves this important issue up to individual developers. Google will be tempted to do that: independent developers will come up with mods to make it easy to add repositories and pollute the main system with stuff that should never be there. Eventually, Google will have to come up with a neat way of installing software just like it happens in their own Android (one folder with everything in it) or Apple (one folder with everything in it) or any other system where software installation actually works.

With Android, they had an easy life because they could dictate exactly what libraries a piece of software should expect and how the applications should play audio etc. (without the OSS/Alsa/Pulseaudio/KDE/GNOME/whatever-else nightmare). They used Java (or Dalvik) to avoid any mess and to make everything more secure. With Google Chrome OS, they will have to make some choices -- and they are going to displease a sizeable portion of the GNU/Linux user base no matter what they choose to do. The choice they made now is great to create buzz, but I don't think it will last. Or, shall I say, I hope it won't.

A possible (and partial) solution to the GNOME/KDE issue is to create "library bundles" so that in order to run GNOME applications you need the "GNOME bundle", for example. Note that this is different to saying that you need libchockablock.so.3.4.2 -- these would be "super-bundles" which would include everything expected by a GNOME application. If then an application needs specific libraries, it could just have them ready in a private "lib" directory in its file system. This is something similar to what Apple does. Again, it's not going to be pretty (applications will look different, will have problems with sound, and so on); but at least it might work.

Back to GNOME and KDE

Whoever said that competition was good, that it was OK for GNU/Linux to have two competing desktop environments, was crazy. The harm done to GNU/Linux was simply immense. When Google decided to build a web-oriented operating system, they ditched them both. With Google Chrome OS, both KDE and GNOME are suddenly less relevant -- and they will become less and less relevant as time goes by.

Goodbye desktop. Hello web applications. Fingers crossed about the necessary transition.

Category: 

Comments

d0ti5's picture
Submitted by d0ti5 on

The thing is, for me, I am not a purist. I don't CARE if something is KDE or Gnome, I just want the best of what ever I am after. This can get pretty weird, as I have duplicate libraries and the gods only know what else in my $/.

But, the key point is really important - who cares? Who REALLY cares about the way things work. What the vast majority of unwashed users want is for the damned thing TO work. You want to come in, turn it on, and then you go do your thing.

If you asked me what I REALLY want, I want to turn on my system with a button. I do not want to sign in, wait for the boot sequence, watch while he takes the time to negotiate with the router, start up all my startup things.

Nope. Button. On. Think television.

I have some base things that I use the system for - so, I would want my email up, my browser at the ready. I want the sound on, Pidgin and Skype and DestroyTwitter ready to feed me nonsense from my friends (or, just one thing for all that). And no waiting.

It is all fun, and I do love playing with the dependencies and all that. But, really, come one. You have a cold, your team humiliates itself on Sunday, your daughter is angsty, you are fighting with the insurance, and all the bills came in today. Do you WANT to mess with things, or do you want to do things. Some days, you just want to do.

The winner in the game will be the one who makes this stuff easy, no-brainer. Of course, the "command line or nothing" people, the ones who DO care, will be outraged. But, you're missing the point. The interwebs and television and communications, they really are converging.

And no one wants to wait - no, not even a minute. We want it now. Heck, we want it to know that we are going to want it in ten minutes, and have it ready for us. We want auto on, ready to go, and no messing around (unless I want to). That is the market that will carry the next system to the top.

ocratato's picture
Submitted by ocratato on

I think that the author misunderstands the whole point of Chrome OS. It is a Web Browser Appliance. It will never run anything but a browser. You will never be able to run GnuCash on Chrome OS, unless you create a version that has a web interface or is built using javascript so that it can run inside the browser.

To go from describing Chrome OS as not using KDE or Gnome to saying that KDE and Gnome are now somehow irrelevant is strange to say he least. Chrome OS no more needs KDE or Gnome than does your set top box or router. It will boot straight into the browser. There is no "desk top"; there is no ability to start multiple programs; there is no need for programs to interact. The only reason for Linux at all is to abstract away the differences in hardware.

I really do not understand why having choices within Linux is considered "crazy". Does anyone complain that there are too many different soft drinks, or too many varieties of bread, or too many different wines? I like to have my Red Hat Enterprise for the backend servers, (or SuSE if I need to interact with Microsoft), Ubuntu for the desktop, Mint for the media machine, and Fedora for my bleeding edge development. It would be annoying to have to pick just one.

If Chrome OS is not embraced by the Linux community it will be because its philosophy is so opposite to the concept of Free Software. It is hard to understand why someone writing for the Free Software Magazine would miss this point. The point of Free Software is that the user has control because she has access to the source code for the applications. With Chrome the user has not only lost that access, they no longer have a copy of the software binary or control of their own data!

In the domain of information, it is actually hard to imagine something more evil than to take away a user's control to the extent that Chrome OS will.

Ryan Cartwright's picture

Hmm perhaps the biggest set of flames you are inviting are from those who do not use either KDE or GNOME. There are of course plenty of other desktop environments and window managers around. ;)

Also you seem to speak as if there is a somebody who could decide who would win the "desktop environment war". As you know - there isn't. The (I think cool) thing about free software options is that the users decide. The best* stuff should rise to the top so when you get two very string competing products it is because there are two very strong - sometimes competing - communities that use and want them. Yes GNU/Linux could have a de-facto DE in the same way that it has a de-facto C compiler or shell environment but with a DE the arguments becomes a little hazy. Choice is good - even where it seems to have shot the underlying products in the foot.

Complaints about packaging, libraries and user-app locations should really be pointed at the distribution providers not an impersonal collection of software. As ocratato says, Chrome OS will be a browser-based environment. For that purpose most of the "apps" used on it will be browser-based web-apps. Libraries become very irrelevant in that case. This is the turnkey solution that is for the netbooks and home PCs of the world. It fits where Android doesn't but I'm not sure it's intended to end up on an OS where development is done or high-end graphic work or music production. It is in those environments that a larger more complex desktop environment fits. Remember Sun's "Java OS"? This coudl well be what that promised to be. Training environments, libraries, internet cafe's, "basic" home users can all benefit from this kind of system.

But in spite of all that, I'd be pretty (nay very) peed off if "somebody" decided for me which desktop environment was best and thus which one I must use. That's one of the reasons I left the proprietary world behind. Choice is only ever any use to me if I get to make my own choices.

cheers
Ryan
--
Equitas IT Solutions - fairness, quality, freedom
http://www.equitasit.co.uk

* of course "best" here usually means most well publicised - as it does in other situations

butter_icing's picture

Whoa, Tony, in one article, you managed to get wrong at least 3 fundamental things, and ranted weirdly about a bunch of others! This is so sad, provided that you are founder and Editor in Chief of Free Software magazine :(.

1. Chrome is NOT GNU/Linux. Android is not GNU/Linux, neither is Chrome. In business term, they have totally different target markets, totally different feature/benefit sets, built with totally different goals in mind. GNU/Linux aims to be a full-fledged system on your computer, regardless if the computer is PC, server, laptop, etc. Chrome is built to be a netbook OS. Trying to matching features between them is like saying, "if your oranges are red, I would have sold them together with my apples! And your oranges would have been way more popular, so please paint them red". Different, see?

2. Even if they are the same, Google would NOT use either KDE or GNOME anyway. First, each of these packages are 200+MB, with hefty requirements on term of hardware (these requirements are on par with Windows and Cocoa; they are a slightly lower, but the margin is not that significant). Remember the market that Google aims at: netbooks, with weak CPU, modest RAM, little hard drive, and feeble graphic cards. GNOME/KDE on there would be like Windows 95, which is ugly as pre-historic. No, Google would not sell that. If Google used an existing technology, it would have been something like XFCE, or even lighter like FWVM or Enlightenment (Samsung invests here!).

3. The existence of multiple window managers are ESSENTIAL to the concept of Free Software. If Free Software gives you virtually no choices, no personalization, no right to vote for suitable programs, how can you call it "free"? Seriously. I use neither, for one, and I do know of many friends who explore different window manager. Furthermore, having one more choices is crucial. For example, Microsoft is invading GNOME with .NET junks. How to respond? Well, switch to KDE :D. Simple and easy. Let's assume that there was only The One GUI, and someone (not necessarily Microsoft; US government would serve just as well) invaded it, what would we do? First panic (yes, people do that with Mono). Then, flame ware. Then, screaming up and down while everyone still use dangerous components, since there is no replacement. Then, replacement will be written (remember the beginning of GNOME? KDE was a bitch back them, eh?). The existence of multiple window managers, thus, is inevitable and necessary.

4. Trying to push out one-program-for-everything is absurd and ignorant. For example, how many file managers are there on Windows? No fewer than 5. True, only one is widely used (Windows Explorer), but yes, there are others. How many music players are there? How many libraries are there? How many programs that do exactly the same thing are there in Windows? True, it has one kernel and GUI subsystem; that counts TWO of unique software. TWO. While there are a trillions of different programs duplicating each other. TWO. What's what that? How can Windows be popular with so much duplication? It is time to rethink the reasons why GNU/Linux have not had more than 10% of the market share. No, it is not about duplication, as any other serious systems have multiple of these. I believe it is more in education, marketing, and software quality. Trying to throw these away to copy a small part of Windows or Mac OS, the least desirable part, is absurd and suicidal, I believe.

Back to Chrome OS. True, it is new, it's from Google, it's excited. Except that it is super dangerous (cloud computing in general is), and it is totally different from GNU/Linux. Can we, thus, stop those weird comparisons? Thank you!

MountainX's picture
Submitted by MountainX on

> Furthermore, having one more choices is crucial. For example, Microsoft is invading GNOME with .NET junks. How to respond? Well, switch to KDE :D. Simple and easy.

I just did that for the very reason you state :)

sgtrock's picture
Submitted by sgtrock on

This has got NOTHING to do with KDE vs. Gnome. You, the founder and editor of the Free Software Magazine, are willing to accept acceding control of your entire computing experience to a single company? Control of all of your data, all choice of what applications you're going to run to manipulate that data, access to all entertainment funneled through a single company? Because that's really what ChromeOS is all about in the long run. You boot it up into a pre-configured browser. Your home page is set, all your mail, online docs, everything runs through a single organization.

Oh, you say you can change the defaults so you're OK with that? Shyeah, like that's going to happen. Google will quietly place enough road blocks in the way that 99.999% of the population won't bother to do so. Their entire business model is based upon leveraging your personal information to generate revenue.

Nope, ChromeOS is a sucker's play. Microsoft and IBM never in their wildest dreams thought they could acheive this level of access and control over the rest of us. I'd rather go back to the bad old days of fighting with dialup modems and X config files before I'll buy a ChromeOS box.

David Sugar's picture

I personally believe that this is about Google's vision to have a thin web client (with all the crucial social and privacy issues enabling a remote vendor to control all your data and the entire computer experience that it brings), and so I think there actually will never be "place" for local applications in that vision, so the question posed here about it's future may be irrelevant. As to whether GNOME and KDE shot each other (and in the process the GNU/Linux desktop) dead, I think that would be entirely worth considering separately from ChromeOS itself.

As to whether ChromeOS is a GNU/Linux, assuming for the moment it has a Linux kernel and core GNU userland, it could well "technically" be so, even if offering such a radically different vision, and one that I do happen to think fundamentally conflicts with the most basic premises of free as in freedom computing. I recall Android is NOT by any definition a GNU/Linux, in that it has an entirely custom software stack including a custom libc.

Sander_Marechal's picture

Remember the market that Google aims at: netbooks, with weak CPU, modest RAM, little hard drive, and feeble graphic cards. GNOME/KDE on there would be like Windows 95, which is ugly as pre-historic.

What? Have you actually ever used a netbook? I have one of the oldest and most feeble netbooks around: an EeePC 900, currently running Ubuntu 9.10. It is fast. Real fast. At some tasks it is even faster than my main desktop (AMD64 3200+ with 2 GB RAM) such as JavaScript execution in Firefox/Vimperator 3.5.

Netbooks are not feeble. The only thing feeble about them is the crappy OS that most come preinstalled with, which is usually Windows or some Linux that has been butchered beyond recognition by the device manufacturer. Replace that with a proper OS and your netbook can do just as much as your desktop or laptop.

accidentalbits's picture

Google's Chrome OS might well be a variant of Linux, but it is not a desktop. And it doesn't even try to be one. Thus I don't see what impact the Gnome-KDE-divide might have had for the design of Chrome OS. The Google team made it pretty clear that they don't intend to allow any other than web applications - in the end Chrome OS is meant for low-end mobile computers used by non-geeks. Google won't be forced to accept executables on the system simply because they don't intend to sell Chrome OS. After all they are a company which monetizes web services. As things stand now, there will be enough room next to Chrome OS for Gnome and KDE -- and Windows and MacOS for that matter. Whether we still have a full-blown desktop system in every household in ten years, on the other hand -- who knows? That depends very much on broadband connectivity and pricing. In my experience, most home users don't need photo editing beyond PicasaWeb for example, or text editing beyond Google Docs. What they need however is portability and mobility - like visiting the family on Thanksgiving, showing them the latest photos and don't even bother to pack the computer or thumbdrive because they could easily access their stuff from a net appliance over there (without downloading anything permanently!).

As for GnuCash, have you tried http://www.mint.com/?

MountainX's picture
Submitted by MountainX on

>As for GnuCash, have you tried http://www.mint.com/?

There is no comparison. Mint lacks essential functionality.

dowdle's picture
Submitted by dowdle on

Are you going to be happy replacing your existing desktop or laptop/netbook computer that has a full functioning OS with web-browsers and hundreds (or possibly thousands) of native apps installed... with an ARM-based netbook with no local storage, no native apps, and only the ability to run a single web-browser, web-apps, storing all data online?

Chrome OS might do well as a netbook-OS alternative to cell phones... but as a usable computer for existing computer users... Chrome OS is going to go over like Coke 2 (aka New Coke) did.

Have you tried it yet? There are VM disk images available that makes it easy. I tried it, and I realize that the current release isn't the final release, but the general consensus is that there isn't much there of interest. Everything Chrome does I can do in my existing web browser... but I have a plethora of things on my existing netbook that I can't do in ChromeOS.

Less is not more. More is more and less is less. Chrome OS is less.

It might be able to boot in a few seconds (if they reach their goals) and my netbook with a full OS might take about 30 seconds to a minute depending on where you stop counting... but with "Suspend to RAM" I don't have to actually boot it much. Waking it up only takes a few seconds.

I just don't see how Chrome OS is going to compete except for on the hardware it comes pre-loaded on where customers will probably be tied to some mobile broadband data plan with little option to switch to anything else.

Chrome OS will probably do like the Chrome browser has... it had quite a bit of pickup shortly after its release but as the months passed it dropped off and it appears it will continue with an insignificant share of the browser marketshare. Of course the competition is good and it'll probably help improve Firefox, Opera, Safari, etc.

So, did Firefox and Internet Explorer shoot themselves dead because Google came out with Chrome browser? :)

Terry Hancock's picture

Corporate imprimature is not a formula for success with free software.

Google's Chrome will be designed specifically to support Google's goals, and they won't be too closely aligned with users or free software developers. So, I seriously doubt that it will displace the community-based KDE and Gnome.

What I expect is that Chrome will become one of the minor/specialized desktop environments -- joining things like XFCE and Enlightenment. Since Google is backing it, it probably won't die out, and some developers will probably "cleave to it" out of respect for Google. But others will be repelled by a system controlled by a single corporation -- it won't be "free enough".

I don't necessarily accept the objection that Chrome is only a platform for a browser -- while that's clearly what it is now, it's very likely that it will be extended (at least, it will if Google releases the source code). But without much broader interest, it's unlikely to become a full desktop environment.

jabjoe's picture
Submitted by jabjoe on

This is a ecosystem of interchangeable parts. You can even swap out the kernel for a different Unix kernel then Linux. This diversity is a great strength because you can build a free Unix out of pretty much any combination you need. One size does not fit all. Even if a single desktop was a good idea, how are you going to enforce it? This is free software, anyone can write a new one, and anyone can fork this one-desktop-to-rule-them-all. Having multiple desktops not only gives competition and diversity of use, but it enforces modularity, abstractions and standards to keep it all working together. Which makes it more robust and flexible. One of the reasons Wine, and Windows itself, have problems running Windows software is that most of the time there is only one implimentation of the API, on one platform, so bugs and quirks go unnoticed and by the time a new implimentation comes along, the people with the source don't care enough to fix it.

Then there is the effect on people, it's not like desktops are that different, being exposed to multiple kinds of desktop enables people to see through to the fundamentals. If there have only ever seen one desktop, they are going to just freak out when they see another with even really looking at how it's much the same. The whole idea of a single Linux desktop can go in the bin marked "broken Windows thinking", along with a stable kernel interface.

Google have their own desktop, because they have different requirements than the existing desktops.

But I see Chrome OS as the return of the thin client, and it's insane in the mainframe, insane in the brain, and it belongs in the 50s. Each time the thin client has tried to return it's failed. I don't see it working better based on HTML and Javascript (though the problem was always control). I would be very surprised if it doesn't fail or mutate more to a standard desktop.

Muir_Bear's picture
Submitted by Muir_Bear on

1) The biggest threat to Linux is itself (a la Pogo). This business of KDE and Gnome believing that their way is the one true way is poisonous. The kernel maintainers need to tell BOTH desktop mobs to support the other's button order. That way a Gnome user will have their KDE apps resemble a Gnome app (e.g. Cancel/Save). Or a KDE user will have the reverse (e.g. Save/Cancel). That was a very rude shock when I migrated from Mozilla (KDE-style) to FireFox (Gnome-style) and had my buttons flip with NO WARNING. I know about the browser hack in the Chrome folder to flip the buttons but the result is extremely ugly.

2) The column talks about how there wouldn't be any local user-installed software. My understanding is that the Chromium browser will be a component of Google Chrome OS. In that case then there is a chance that Python extensions to the browser would be allowed.

3) [Re : jabjoe + Thin Clients] I'm a former mainframer from the 1970's and 1980's. Glass teletypes were and still are the pits. But 3270-type gear with some local storage worked fine back then and is still around in places like insurance offices. Yes, thin client hardware is a niche product but it's like edged weapons. Do you want to try and eat a meal with the help of a claymore or a claybeg ? Or would you prefer a dirk ?

A netbook-like rig with Chrome OS gains reliability and portability for giving up local packages. My own laptop when fully kitted out weighs in at around fifteen pounds (15 lbs.) in a briefcase-like bag. That's in addition to the feldtasche* I usually carry. If I could leave that baby boat anchor at home for most of my runs to a Wi-Fi hotspot I would. There's room to spare in my feldtasche along side a book to read on the bus or train.

Edged weapons :
claymore - broadsword, 55"+
claybeg - saber, 36" - 40"
dirk - dagger, 6" - 12", sometimes longer

* feldtasche, from German, field bag
e.g. a medium-size canvas tote for food, books, and other loose junk

Author information

Tony Mobily's picture

Biography

Tony is the founder and the Editor In Chief of Free Software Magazine