Web 2.0 Summit: The Browser Is What Matters (InformationWeek) [LWN.net]

InformationWeek covers comments by Google's VP of product management, Sundar Pichai, at the Web 2.0 Summit. "To the suggestion that Chrome OS -- the operating system that Google is developing around its Chrome browser -- is on a collision course with Windows, Pichai responded that the world is entering a period of tremendous innovation in personal computing. "Browsers are suddenly hot again and I think operating systems are too," he said, referring both to Chrome OS and Android, Google's operating system for mobile devices "There haven't been other choices for a long time," he said. "Most operating systems today were designed before the Web existed." "The goal with both our efforts is to get great free open source software stacks out there," he said. In the case of Chrome OS, everything is built around the browser."

to post comments

Web 2.0 Summit: The Browser Is What Matters (InformationWeek)

Posted Oct 22, 2009 22:03 UTC (Thu) by brinkmd (guest, #45122) [Link] (14 responses)

WRT this: "There haven't been other choices for a long time. Most operating systems today were designed before the Web existed."

That's simply ridiculous.

1962 Network design on paper
1969 Arpanet runs with 4 nodes.
1969 Early versions of Unix developed
1973 TCP/IP design starts
1973 Unix rewritten in C
1975 First TCP/IP test network
1981 Early versions of Windows
1983 Arpanet becomes part of the internet
1990 Windows 3.0
etc

It's fair to say that some operating systems were written before the Web, the most often used OSs were written while the web was developed and *all other* operating systems since then were written while the web existed.

There have been plenty of operating systems all the time, every year probably dozens of research systems are added to the list. But few of them have the Google money behind it to push it.

The more important quote from the article is:

"As a user, you don't install software, you don't maintain software," said Pichai. [...] "The browser is what matters," he said.

This just means that the Chrome OS is Google's way of making sure that you can only run Google's web applications. The thin client is back.

Web 2.0 Summit: The Browser Is What Matters (InformationWeek)

Posted Oct 22, 2009 22:56 UTC (Thu) by donbarry (guest, #10485) [Link] (8 responses)

And worse, that Google can "sell" modified GPL software
for which they do not have to release source because they are
not "distributing." The thin-client paradigm, applied to the
web as a whole, negates many of the freedom protections of the
GPLv2. Little software as of yet uses GPLv3 with or without
the Affero clause. I suspect that will change as people see
the erosion of freedom in this brave new model. The question
is then, too little, too late?

Web 2.0 Summit: The Browser Is What Matters (InformationWeek)

Posted Oct 23, 2009 2:23 UTC (Fri) by drag (guest, #31333) [Link] (7 responses)

I wouldn't worry about it to much. The idea that web-based applications are
going to displace host-based applications is laughable at best.

Besides the obvious and severe issues of just plain usability, flexibility,
and performance issues that web-based applications severely lack you run into
the fundamental reality that web-based applications don't scale as well as
host-based. That is as your software becomes more popular the cost of web-
based stuff goes up massively. Client-side scripting (javascript) and content
delivery networks can help mitigate the problem, but its a fundamental
reality that pushing the workload out to the individual end nodes is just
fundamental concept for scalable applications.

Yup

Posted Oct 23, 2009 7:50 UTC (Fri) by khim (subscriber, #9252) [Link] (4 responses)

Client-side scripting (javascript) and content delivery networks can help mitigate the problem, but its a fundamental reality that pushing the workload out to the individual end nodes is just fundamental concept for scalable applications.

That's why Google makes this possible via browser. It's as simple as 1, 2, 3. You can be sure it's only a beginning: if you view the browser as an OS... you only need "drivers" for your hardware - and your browser app will have the same capability as native one.

It's not exactly "thin client" as it was envisioned before: software can work on client side if it's needed, but web site controls it when it's connected to the internet.

Time will tell if the idea will fly or not, but to dismiss it just because "browser just can not do X or Y" is ridiculous...

Yup

Posted Oct 23, 2009 15:25 UTC (Fri) by drag (guest, #31333) [Link] (3 responses)

If your saying that people will end up writing local applications in a
combination of Javascript and HTML then that is certainly possible and is
likely to succeed. It is really fundamentally no different then writing
your applications in C# and WPF or Python and Glade. If you think about it
then in that situation then the "browser OS" is no thinner or lighter then
if I write a Python or Mono application and give it the ability to connect
to some external data source on the internet.

Otherwise thin client concept is much much older then actually having your
own PC and having your own copy of the software running on your own PC. And
so far in just about every case when people are given the choice between
depending on local storage and local environment for the majority of their
computing tasks versus going thin people have always chosen to go with the
"thick client" type setup, except in very specific circumstances.

What protocol your trying to do the thin client over is kinda irrelevant. I
don't think that there will be any paradigm shift simply because your using
HTTP instead of X11 or RS232.

It's not about developers, it's about users...

Posted Oct 23, 2009 16:36 UTC (Fri) by khim (subscriber, #9252) [Link] (2 responses)

It is really fundamentally no different then writing your applications in C# and WPF or Python and Glade.

Yeah. There are small differences, but they are not so big.

If you think about it then in that situation then the "browser OS" is no thinner or lighter then if I write a Python or Mono application and give it the ability to connect to some external data source on the internet.

If you only think about resources needed then yes. But the fact is: even current "underpowered" ntebooks have many GHz CPUs and gigabytes of RAM. The story here is not about development model but about deployment model. You don't install the software, you just visit the webpage and start to use it. If you sync often enough then you can just swap your netbook with newer model and continue from where you left!

Actually this model was invented by Microsoft and it's why MSIE just refuses to die: a lot of corporate "webapps" are just plain old Windows Apps embedded in ActiveX! Some even talk to server via proprietary protocol, not via HTTP! The problem here is the fact that Windows is just not designed for this sort of usage: even if you only plan to use browser and some "webapps" people install all sort of junk - and it's very hard to forbid this abuse and still keep working system (since "webapps" are just plain old Windows apps you can not just forbid to install and run Windows apps).

Google plans to start from scratch and forbid any kind of model abuse from the start. Easy to do, but will the developers and users accept it? Time will tell.

It's not about developers, it's about users...

Posted Oct 23, 2009 16:44 UTC (Fri) by drag (guest, #31333) [Link] (1 responses)

So..

Fundamentally what your saying is that the difference between the "OS" and
the "Browser OS" in the way Google Chrome is going about it is that with a
traditional operating system people have to run a installer while with
"Chrome OS" if they plug a URL into their OS then the application is
automatically downloaded and executed in a sandbox.

If so then that is a somewhat interesting way to work around how much it
sucks for a regular person to install software in Linux.

That's the idea...

Posted Oct 23, 2009 17:02 UTC (Fri) by khim (subscriber, #9252) [Link]

With "Chrome OS" if they plug a URL into their OS then the application is automatically downloaded and executed in a sandbox.

Yup. That's the beauty: you can use you app on any popular OS (Windows, Mac, Linux) and on dedicated OS (Chrome OS) too. I'm not even sure there are plans to allow any apps not usable in regular Chrome.

Web 2.0 Summit: The Browser Is What Matters (InformationWeek)

Posted Oct 23, 2009 15:14 UTC (Fri) by Hanno (guest, #41730) [Link] (1 responses)

I wouldn't worry about it to much. The idea that web-based applications are going to displace host-based applications is laughable at best.

Google Apps has replaced own infrastructure for many folks I personally know, both private groups as well as small and medium companies. They are happy with it.

So I disagree with your arguments against web-apps. The one reason I hate them is that my data is stored on Google servers. My colleagues and I use Google Calendar a lot. It is hands-down the best calendar app so far - and I try to replace it with a local host-based app at least once every year, but failed every time.

web-based applications don't scale as well as host-based

You should talk about that with a company which I think is called Google. I hear they know a bit about salability and delivering web applications to millions of users.

Web 2.0 Summit: The Browser Is What Matters (InformationWeek)

Posted Oct 23, 2009 16:34 UTC (Fri) by drag (guest, #31333) [Link]

Yes Google is a monster.

But if your talking about something that will get widely used then we are
not talking about millions of users, but billions. It is estimated that
there are now over 1 billion computers running Microsoft Windows. And this
install base is usefully providing a platform to run people's software that
can barely afford to own and opperate 3-4 servers, much less a cloud system
that can offer services to thousands of people, much less hundreds of
millions of people.

I can write a piece of software, say a video game, and throw that
application on a web server somewhere for people to download.

They download it, like it, and then lots and lots of people decide they
like it. For each user that downloads it then all that it costs me
personally is the bandwidth it takes to download it. And if it is popular I
could distribute it under bittorrent and then the cost is even a fraction
of that.

The difference between hundreds of users vs hundreds of thousands of users
is not that big of a one. Now imagine if the video game required some cloud
service like Google's or Amazons and I had to rent out server time
proportional with the demand... then that would be massively expensive.

Even if I was doing a MMORPG were people log into a server it would still
be vastly cheaper for me to provide a single download were the vast
majority of activity happens on the client machines and my online server
just manages relaying the protocols that make up the online interactions
and stores some trivial data in the form of scores or a avitar's
possessions.

Web 2.0 Summit: The Browser Is What Matters (InformationWeek)

Posted Oct 22, 2009 23:32 UTC (Thu) by epa (subscriber, #39769) [Link]

One event is clearly missing from your timeline:

1991 First public announcement of the World Wide Web

And yes, the typical OS design, whether you refer to the high-level interface of mouse and windows or the low-level guts of processes and files, does indeed date to long before the Web existed, let alone when it became widespread.

Web 2.0 Summit: The Browser Is What Matters (InformationWeek)

Posted Oct 23, 2009 5:15 UTC (Fri) by Arker (guest, #14205) [Link] (3 responses)

"The Web" is not the same as "the Internet."

The Internet existed and was useful and used LONG before the web appeared.

Web 2.0 Summit: The Browser Is What Matters (InformationWeek)

Posted Oct 23, 2009 10:49 UTC (Fri) by brinkmd (guest, #45122) [Link] (2 responses)

Sure, but what's the web in your opinion?

I challenge anybody to come up with *any* reasonable definitions for "OS" and "Web" such that the following sentence becomes anything else but marketing hyperbole:

"There haven't been other choices for a long time. Most operating systems today were designed before the Web existed."

What he actually wanted to say is: "Most operating systems today were designed before Google's business model existed." And even that is wrong, as there surely have been operating systems for thin clients that receive advertisement supported content before.

Web 2.0 Summit: The Browser Is What Matters (InformationWeek)

Posted Oct 23, 2009 15:33 UTC (Fri) by rfunk (subscriber, #4054) [Link] (1 responses)

The web is the cross-site hyperlinked system based on HTTP+HTML, which
began in 1991 but didn't really begin to take off until Mosaic added inline
images in 1993. Then businesses (far beyond Google) brought the web to the
masses starting around 1998.

- Unix was designed in the 70s and 80s. Linux's design was heavily based on
that of Unix, and began in 1991.

- Windows NT (the basis for modern Windows) was designed in the late 80s
and early 90s, based heavily on the older VMS OS (designed in the 70s and
early 80s). And Microsoft was notoriously late in noticing the importance
of the web.

- Mac OS X was designed in the late 90s, based heavily on NeXTStep
(designed in the early 90s), which in turn was based heavily on BSD Unix
(early/mid 80s).

Web 2.0 Summit: The Browser Is What Matters (InformationWeek)

Posted Oct 23, 2009 17:01 UTC (Fri) by drag (guest, #31333) [Link]

Yes your right. Hilariously right.

Web is indeed just the international network of interlinked websites. HTTP
= Hypertext transmission protocol.. that a protocol designed to transmit
text with links to other text instead of references at the bottom of the
page. Infinitely more useful way to document and share knowledge and
science then a bunch of books and libraries.

A fundamentally simple concept yet depended entirely on the REST concept to
scale and function properly. REST is "Representational State Transfer" and
is fundamental to how the web is able to function properly. Client
communicates request to server. Server processes the request and sends the
data back to the client. The state of the website is stored in the client
browser which is almost always at rest and consumes almost no resources on
either the client nor the server. The client controls the state and
transmits to the server when it is time to change the state.

All in all it is a very elegant system and deceivingly simple. Violations
of this concept is now why it takes a average of 1-2gigs of RAM to run a
LAMP server and visiting certain websites can bring my netbook to it's
knees.

Of course the internet itself only became publicly available in 1989. :)
Which is just a few years before WWW became realized.

Before that you had proprietary networks and BBS's. I did not get involved
in networking until the Internet existed so I missed out on all the fun of
calling other people's computers over POTS and such.. but I did get to see
what it was like before Mosaic when I used Kermit and Gopher to connect to
the internet.

Web 2.0 Summit: The Browser Is What Matters (InformationWeek)

Posted Oct 22, 2009 23:54 UTC (Thu) by elanthis (guest, #6227) [Link] (9 responses)

Unless Cheome OS focuses on solid highend graphics, it's dead in the water. Nobody
except Internet businesses and Linux nerds think that the browser is all people use. The
vast majority of computer play games. Real, quality, triple A games, not TuxRacer. Kids,
college students, and even many of the "younger professionals" out play at least a few
games. The entire drive the industry has in more powerful consumer hardware focuses
around games. Much of the iPhones success is from it's gaming ability. Until a Linux based
device is capable of satisfying the needs of modern games, it will never sarisfy the needs of
real people.

Web 2.0 Summit: The Browser Is What Matters (InformationWeek)

Posted Oct 23, 2009 0:14 UTC (Fri) by ewan (guest, #5533) [Link]

Of course, a lot of 'real people' have got heartily sick of the unreliability and anti-piracy hoop-jumping of PC gaming and simply split their needs over two devices, a PC for general purpose non-gaming applications, and a console for games. For those people the PC really doesn't need to run games.

Web 2.0 Summit: The Browser Is What Matters (InformationWeek)

Posted Oct 23, 2009 3:53 UTC (Fri) by eru (subscriber, #2753) [Link] (4 responses)

Nobody except Internet businesses and Linux nerds think that the browser is all people use.

The funny thing is, I could say almost exactly the opposite, based on observing my non-nerd friends and relatives! They spend their computer time in the browser almost exclusively, except for some small wordprocessing stuff, and a browser-based app could handle that, too. None of them see games as a particularly important application for computers. Most of them don't play any computer games at all.

Web 2.0 Summit: The Browser Is What Matters (InformationWeek)

Posted Oct 23, 2009 14:21 UTC (Fri) by allesfresser (subscriber, #216) [Link]

I second eru's comment. Games are unimportant to the vast majority of the people I interact with. But perhaps it's a generation gap, as it so happens that I mostly interact with people over the age of 25.

I also second that comment

Posted Oct 23, 2009 15:38 UTC (Fri) by dbruce (guest, #57948) [Link] (1 responses)

People who spend a lot of time with computers and software get a distorted picture of what the general public really does. The browser really *is* what the majority of casual computer users spend nearly all their time in. I would guess that less than 50% of the population has ever installed any software application, less than 10% has opened the case to replace any hardware, and less than 1% has ever installed an operating system.

AFAICT, people usually buy computers and use whatever hardware and software comes with the system. If anything needs to be changed, they either pay the service department at the local appliance store to do it for them, or drag in the unlucky family member who serves as the "computer guy" (someone like the graphics-requiring game players mentioned above).

I also second that comment

Posted Oct 25, 2009 11:59 UTC (Sun) by oak (guest, #2786) [Link]

And almost all the games they play are something running in Flashplayer.
There are sites dedicates to Flash games (mostly for teenagers I guess)
and social web sites like Facebook also provide (over-the-net multiplayer
= social) Flash games.

Triple-A (especially single player) games sound more like a nerd thing
like all non-casual gaming.

Web 2.0 Summit: The Browser Is What Matters (InformationWeek)

Posted Oct 29, 2009 9:42 UTC (Thu) by stevan (guest, #4342) [Link]

Yet one of the accusations against greater mass uptake of Linux as a standard OS is "lack of
applications." It seems that a lot of the cloud offerings are even more selective about app
availability, yet there is an assumption that this aspect is no longer the issue it is seen to be,
in a slightly different context. I don't see how this can be both true and false.

If, on the other hand, the proposition is a greater variety of productivity sources, that makes
greater sense, but western business doesn't seem to see things like that. <ABBA> The
winner takes it aaaaaaaaaaal </ABBA>

Web 2.0 Summit: The Browser Is What Matters (InformationWeek)

Posted Oct 23, 2009 4:14 UTC (Fri) by cpeterso (guest, #305) [Link]

Is Chrome OS trying to be anything more than a tablet/netbook OS?

Web 2.0 Summit: The Browser Is What Matters (InformationWeek)

Posted Oct 23, 2009 9:20 UTC (Fri) by Cato (guest, #7643) [Link]

Google is proposing O3D for high performance 3D graphics on the web: http://code.google.com/apis/o3d/

This will be accessible both from JavaScript and Google Native Client: http://www.vizworld.com/2009/06/google-welds-native-clien... and http://ostatic.com/blog/google-forging-ahead-with-native-...

Since you can already play Quake using a browser plugin, I see no reason why O3D or Native Client wouldn't work for this. The issue will be uptake by game developers, but presumably once you develop for this sort of 3D / native code API, operating system differences become irrelevant so I can see this being attractive to developers as gaming moves to the web in any case.

Web 2.0 Summit: The Browser Is What Matters (InformationWeek)

Posted Oct 23, 2009 15:33 UTC (Fri) by Hanno (guest, #41730) [Link]

Nobody except Internet businesses and Linux nerds think that the browser is all people use.

The folks you hang around with may be hardcore gamers, but you should not generalize from the little subset of the population you know. Real people actually do use little more than a wordprocessor, a browser and an email program. My mum asks me about home accounting and text editing, but not about triple A games and solid highend graphics.

The best selling graphics card of the PC business is the chipset graphics of Intel - very very lowend graphics.

And the iPhone isn't known for its highend gaming ability and wasn't advertised as a portable gaming console when it entered the market. Its games are not triple A, but casual games, often with very poor graphics.

The Nintendo DS has won over the PSP and the Wii over the PS3. Seems that highend graphics isn't all that counts.

Web 2.0 Summit: The Browser Is What Matters (InformationWeek)

Posted Oct 23, 2009 2:36 UTC (Fri) by qg6te2 (guest, #52587) [Link] (3 responses)

Given that Google is "now working on preparing a 2.6.26-based kernel for deployment in the near future" [1], I sincerely hope that for Chrome OS the smart people at Google decide to use a newer kernel than this ancient relic.

I'm worried as I have first hand experience with large organisations making technical decisions based purely on bean-counter advice, whereby engineering efforts are deliberately constrained in order to "share as many common parts as possible" (read: reduce costs). This resulted in, to put it mildly, a reduced end-user experience.

Google's recent revelations (seemingly regarding their data centre machines) include stuff such as "forward-port[ing] the old O(1) scheduler", which would be plain crazy for a desktop system. Using newer kernels is also advised due to KMS functionality.

[1] http://lwn.net/Articles/357658

Web 2.0 Summit: The Browser Is What Matters (InformationWeek)

Posted Oct 23, 2009 3:33 UTC (Fri) by csamuel (✭ supporter ✭, #2624) [Link] (1 responses)

2.6.26 is what they use internally, Android devices seem to run 2.6.29.

However, they don't seem to be showing any interest in their Android code
in the mainline kernel leading to Greg K-H threatening to drop both
Google's and Microsoft's code from the staging tree in 2.6.32.

Web 2.0 Summit: The Browser Is What Matters (InformationWeek)

Posted Oct 23, 2009 4:41 UTC (Fri) by qg6te2 (guest, #52587) [Link]

they don't seem to be showing any interest in their Android code in the mainline kernel

Pretty slack for an organisation which has gained so much through the use of Linux. It's not like they have a shortage of engineers. Perhaps they've conveniently interpreted their "do no evil" motto as "do nothing".

Web 2.0 Summit: The Browser Is What Matters (InformationWeek)

Posted Oct 23, 2009 15:17 UTC (Fri) by rfunk (subscriber, #4054) [Link]

The plans you referenced are for their servers, and have nothing to do with
ChromeOS (or Android).

Web 2.0 Summit: The Browser Is What Matters (InformationWeek)

Posted Oct 23, 2009 10:54 UTC (Fri) by tnoo (subscriber, #20427) [Link]

Ironically, now that the SUN slogan "the network is the computer" is
becoming reality, that company went out of business.

Web 2.0 Summit: The Browser Is What Matters (InformationWeek)

Posted Oct 23, 2009 17:42 UTC (Fri) by clugstj (subscriber, #4020) [Link]

WRT this: "Most operating systems today were designed before the Web existed."

So what. How is this [fact?] relevant. I guess when you listen to a VP of anything, you have to expect to get lots of words with little content.

Web 2.0 Summit: The Browser Is What Matters (InformationWeek)

Posted Oct 23, 2009 19:44 UTC (Fri) by dkite (guest, #4577) [Link]

Everything is wonderful compared to something else. Compared to the
software stack from Microsoft, and all the hassles, cost and management
issues, what google is offering is wonderful. Just being free from
proprietary lock-in is amazingly productive.

But once you get past that, anything beyond simple usage for browser based
apps sucks. Latency, the limitations of the platform, simple stuff like
focus and popups, notifications, etc. services that a rich desktop can
offer are awful on the browser.

So google needs to build something thicker. They can't kill their bread and
butter revenue streams by writing an application that does something
better, they don't have any to kill. So they have a chance at doing
something interesting.

Kinda like Microsoft when they changed the business by offering a suite of
applications.

Derek