Friday, July 10, 2009

Chrome OS and Google World Domination

The web platform is starting to get annoying in its attempt to re-invent and re-create everything that previously existed in the Desktop space.

Finally, Desktop developers are beginning to wake up to this problem.

For years, browser-developing companies have tried to turn the Browser into an operating-system-like autonomous system. While the rest of us were trying to figure out how to get our favorite tools (Python) into the browser space.

No one seems to have fully realized yet, that it is much easier for us, Desktop denizens to conquer the world than for the browser people. They miss almost everything from the kernel up, while all we need to do is integrate better the web into the desktop.

We don't even need browsers. Minimize your browser (and all other windows you may have open) an look at that beautiful picture in you desktop. When was the last time you had a productive relationship with your desktop? Mine is used mainly to hang a pretty picture and place a bunch of useless files stashed to one side so they don't mess up the view. But even with my beautiful wallpaper unblocked by the icons on the desktop, I rarely look at it, because my browser is always blocking the view.

IMHO, the web should take the place of the walpaper and desktop apps should be gradually adapted to be more and more web aware, so that we don't need thousands of browser extensions duplicating the functionalities of our desktop apps which can't deal with the web.

My extension loaded Firefox, currently takes almost as long to start as Ubuntu to boot. Why do I need this mamooth when I could have the web on my desktop and integrated with all of my operating system? Why on earth do we need yet another OS?

Chrome OS according to recently released information, is a linux kernel with a light-weight graphics layer (not X) dedicate to a browser (chrome in this case). They are doing exactly what I proposed above, but I am not prepared to give away my desktop for that. What we need to do now, is to prove that the browser is the problem, not the operating system.


antoniofonseca said...

I think you're completely wrong, because what
you propose is just what Microsoft has been devoted to
make throughout the 1990s with the integration of the IE and

As you can see this is nothing new or even been proven effective and safe.

antoniofonseca said...

And a think that Google is trying to implement something really new.

Paddy3118 said...

Microsoft says integrate the browser into the OS. Google says integrate an OS into a browser. The ideas are the same same, what is different is who ends up in control.

Hugh Isaacs II the Maître d' said...

I have to disagree with you.

One thing to note about Chrome OS is that it's coming out later this year and will be able to support all apps that Chrome will support.

Chrome will soon have Googles Native Client plugin built right into the browser, enabling native applications (e.g. Python) along with the fact that Gears adds support for offline applications and installing a web app to your desktop.

For you to state that you want the web to be more integrated into the desktop is basically you asking for a web browser with support for native applications, there's not much going for the desktop except that.

Flavio Coelho said...

@antoniofonseca: What microsoft has been doing is simply shipping the OS with a browser pre-installed, so people won't need to install competing browsers.

The browser was never truly integrated into the environment, it has always been just another application.

Flavio Coelho said...

@Hugh: Support for native apps from within the browser will never be as efficient than running the same apps straight in the OS because it is adding an unnecessary extra layer between the hardware and the application.

All current OSs need is to do away with the desktop (as in desk top) metaphor and put web content right in its place.

hcarvalhoalves said...

Disagree with your post.

The "browser" is actually a good platform that brings something desktop systems weren't capable of giving for decades now: cross-compatibility and standards.

You can't write a single software that will run on Windows, Mac, Linux, Symbian, iPhone and Android. But you can indeed write a ubiquitous web app that all those platform browsers can access. Not to meant that those applications are effectively sandboxed on the remote server, thus enabling transparent upgrades and more security than if it were running on the client.

Also, when more client-side functionality is needed, there are plugins (like 3D support in Flash) and upcoming standards (like the "video" and "audio" tags on HTML5) to help iron this out.

In the end, I think the situation is exactly the opposite: the desktop market only works well when you have a single standard to support (like it was in the past decades with Windows). Now that it's becoming more and more heterogeneous, with new OSes and different devices, you just can't win. The best common denominator you have is the browser, so people are going to it.

Michael Foord said...

The main problem with the browser as it currently exists is that it just *can't* do a tonne of stuff we do with desktop apps.

Store your music in the cloud for playing offline?

Same with videos.

Software development in the browser? Possible but currently a horrible experience.

Offline access to *all* your documents - possible but currently very sketchy.

No web interface to skype.

Browser access to your webcam and microphone - flash I guess but not Javascript.

Computers have moved a long way from the dumb terminal and mainframe model, and for good reason - it was pretty awful as a user experience. Why some people are hell bent on returning is a mystery to me.

Flavio Coelho said...

we have plenty of crossplatform alternatives rightnow, without the HTML+Javascript: Java, .NET, Python and all other virtual-machine based programming languages. We even have Gui-toolkits which are cross platform: Qt is the best of them IMHO. Also, if you stick to opensource libraries like Qt or GTK you can write C and C++ apps which are perfectly cross-compilable.

Paddy3118 said...

Whither Python?
If the OS is just enough to run the browser, will we be left writing everything in 'browser supported languages'?

j_king said...


Using the web as an interface was a work around when developing cross-platform applications was a real problem.

The browser is a poor runtime for application interfaces. It's stateless. Every browser interprets the "standards" of HTML/CSS/JS differently. They all implement their own idea of accessibility differently. The fact that these web applications work at all is really amazing considering the sheer number of layers and serialization steps a simple bit of data has to go through before the user sees it. The amount of technical proficiency required to deliver the more sophisticated web interface is increasing dramatically to the point where tools like GWT and teams of engineers need to provide high-level abstractions to keep us hard-working grunts from losing our minds.

And the funny thing is, these web applications are mostly re-hashing years of hard work that went into cross-platform desktop development.

Is running in a browser over an Internet connection a feature in and of itself?


However, Google is likely to see the world through rose-coloured glasses and wouldn't know the difference. With a hammer in hand, every problem looks like a nail.

Transparent upgrades are something else altogether. I grant that updating desktop applications is difficult because even Linux-based distro's can't seem to agree on a single packaging and distribution system and proprietary vendors have barely even caught on to the idea. However, if that were a non-issue it would take less expertise to update a desktop application than a web application in the long run. As one adds more users to a web application, one requires more expertise in maintaining and updating that system. Local resources are easier to manage since there's nothing to scale. Servers could then be reduced to simply storing data and managing access to it statefully and avoid a lot of hassles.

But alas, such ideas are hardly new and definitely not on the bandwagon.

As for Google Chrome OS: big deal. The only draw is if you're an existing Google apps user. Now you can drop some cash on yet another electronic doodad so that you can use their apps... wait a minute... what's the benefit again? NOTHING?