Feeds

back to article Google Chrome beta turns on native code machine

With its latest Chrome beta, Google has turned on Native Client, its rather bold effort to securely run native applications inside the browser. This means that Native Client is slated to make its official debut with Chrome 14 in September. In a recent interview with The Register, Google vice president of engineering Linus Upson …

COMMENTS

This topic is closed for new posts.
Anonymous Coward

OS by stealth

If someone said ten years ago that Google would undermine MS Windows as an OS by creating a web browser that supports <peek> and <poke> tags (old timers will get that joke, and its irony), use web languages as an application platform and allow websites to run C++, they'd have been locked up in a padded cell and sedated.

12
0
Alert

Get the Popcorn, this is going to be good.

I guess we'll see if Google's sandbox is up to the inevitable attempts to break out of it.

6
0
Anonymous Coward

I guess...

I guess that's one of the reasons that Google have to approve and supply the programs that people write for Chrome. Security through not letting anyone actually try to attack the platform, without prior approval. It's not really security, but it seems to be working for Apple with the iOS.

0
0
JDX
Gold badge

What's the point?

A browser-specific extension in a minority-use browser... it might be really cool but who is going to target it when it means their stuff only runs on Chrome?

5
0

Chrome frame

If the user is conditioned to accept all plug-in installation requests (or anything witha Yes/No dialog), the computer probably already has Chrome frame installed. So the penetration may be close to IE levels.

Of course, nobody ever targeted ActiveX when it meant their stuff only runs on IE. Right?

0
0
Bronze badge

I'll point to it

If the minority who use chrome are geeks, what percentage of them do you think are administrators? In my estimation, I'd guess 90%. What percentage of administrators use web interfaces? I'd guess 99%. If they can break out of the sandbox, they've probably got it made.

1
1
Devil

Well, it depends on the apps

The availability of Kindle has just made me install it.

One app at a time... Even if that app is HTML5+offline data and not native code.

Does Mozilla like it or not, but it will have difficulty playing catch-up to a platform with the Chrome development budget.

0
0

!minority

More people use Chrome than FireFox, apperently...

0
0
Gold badge
WTF?

Re: Chrome frame

Hmm, one problem with that theory. I've never even *seen* a dialogue asking for permission to install Chrome frame, let alone decided to allow it.

What is there that requires it?

0
0
Happy

Arse about face

Dalvik <--> PNaCl

There, fixed it for them.

0
0

The point is Chromium

A much more secure OS bound to a much more secure browser. Native code simply opens the door to a larger developer community and a larger application base.

1
0

Title shmitle

Wait a minute, the point of Native Client, as they say, is to allow applications that require faster performance to execute natively and thus gain speeds not possible with JavaScript, right?

And then we see this,

>> "Upson told us that when released, PNaCl will offer speeds comparable to JavaScript."

So... What's the point?

dZ.

6
1
Facepalm

Not turned on by this

As a web developer I'm not turned on by this. Why are they doing this, and what is the appeal to developers? And anyone who spends time developing C++ Pepper stuff has a limited audience of Chrome users, right? What's the point?

Google have employed too many developers who are struggling to find useful things to do.

2
0
Bronze badge

web developer not turned on by

Of course not. Writing web apps in C or C++ won't make them any faster, smarter, or easier to develop.

But the software developers in Google didn't write Google in Java Script, and they want to write web apps in the languages with which they are familier.

0
0
Silver badge
Holmes

It's All About the APPs stoopid!

This native code malarkey isn't intended for web devs to put on their general access web sites, it is intended to allow developers targeting ChromeOS to deliver apps that run at at so-called "native speed" to Chrome users via the Chrome App store.

0
0
Facepalm

Cognitive Dissonace

So let me get this straight - the same people who for years were complaining how Microsoft put all these extensions into its browser and that they should just stuck with the HTML spec are now cheering for Google putting extensions into their browser and not sticking with the HTML spec.

14
1
Silver badge

NaCl is such a waste of time

NaCl executes programs which are written with native instructions (e.g. x86) sandboxed with hacks which not all CPUs may support properly.

It boggles the mind Google are even bothering with this considering they're supposedly working on a Portable NaCl which uses LLVM so the same program runs anywhere.

Google are muddying their own waters here and they'll be lumbered with this silly stopgap for years to come.

0
0
Trollface

PNaCl, I'll take that with a pinch of salt!

As per title.

4
0
Devil

Will you load that pinch of salt onto RDX cartridges too?

No boom today.... Boom tomorrow... There will always be a boom tomorrow...

0
0
Silver badge
Headmaster

This looks like a pile-up of several concerns...

1) A Virtual Machine with its own bytecode (not bitcode) instruction set (One would have hoped the 'compiled code' would just be the parse tree result, which gives you lots of flexibility, but noo...)

2) A sandboxing implementation

3) A way of porting of "native code" between machines (Why?! People compiling to native should get what they deserve - either high performance or utter irrelevance.)

4) All of the above packed into a "browser"

Sounds like extreme complexity about to give people bad hair days.

0
1
Ru
Silver badge
Meh

Its the 'inner platform' effect

Turns out that in order to replace a general purpose operating system, you need something that works an awful lot like a genera purpose operating system. Only with all the libraries and UI bits rewritten. And requiring your developers to jump through yet another set of hoops in order to run their applications on your platform.

Meet Boss 2.0. Same as the old boss, but shiner, slower, less accomodating and a whole lot more contrived.

1
0

(untitled)

Don't use it.

A proprietary internet, beholding to one browser has only recently been brought back to open standards.

(Die IE6)

Use Opera, Firefox, IE8 / IE9.

The Advertmongers that are Google Inc only give you free stuff to bribe you to give them data. Your data is worth far more than a sweetie (candy for you Americans). The personal data that is inadvertantly given to Google is far more than is needed for them to build a profile of you AND with the streetview IP address snaggling they KNOW where you live too.

Do you use Android? Registered your phone? What's the betting they know your name too?

Information is power. (to paraphrase the old quotation) Google has enough on you to model your behaviors.

Google has power, too much of it. I believe that "do no evil" died many moons ago.

6
2
FAIL

MY NAME?!

What's the betting they know my name?

I'd say those are pretty good odds, as I provided them with my name when I signed up for a Google account.

0
1
Ru
Silver badge

"Your data is worth far more than a sweetie"

Is it?

I was always under the impression that google was after ad impressions. It is so far the only thing they can do well which also generates them serious amounts of revenue. Every single other thing they've ever done and tried to make money from has been a disappointment, either to consumers or google's bottom line.

They wouldn't know what to do with your data. They just want you to click the ads.

0
1

wait I've got an idea

What about native apps targeted directly to the OS, but all in a database on the internet. Packages installed from this db are all kept upto date with the db. Everything is signed and there is a strong trust system for db write access. You can save and restore the selection of packages installed. Oh wait, we've had that for years!

3
0
WTF?

HTML5 anyone?

This sounds like an attempt to create a technology for writing 'apps for the web' rather than using the web it's self - no thanks. Let HTML5 mature and avoid this nonsense. This is simply an attempt to introduce native apps to the cromebooks in retrospect.

2
0
Ru
Silver badge

From another point of view

it sounds like a way to develop apps for the browser-only future that doesn't require you to use bloody javascript. I do not look forward to the day when it becomes a necessity to use this awful language for writing any and all applications.

1
0
Silver badge

Javascript awful?

It has some terrible features. But it's also incredibly powerful and flexible. (I've been coding in it for the past year or so and it's very different to working with strongly typed languages, but that doesn't make it bad!)

Haven't you heard of Node.js?

I suggest reading "Javascript: The Good Parts" by Douglas Crockford. Should explain to you why your "awful" comment is just plain wrong. Of course javascript isn't suitable for writing every app, but there's no language that's good for everything, that's why there's so many of them!

0
0
Silver badge

Surely this is a bollocks idea?

Need I say more?

2
1
Bronze badge
Facepalm

wow, back to the future with segment registers

25 years ago Intel made it possible to change segment size (from 64kb to 4gb) and made the segment registers redundant, and now a microcode design choice designed for the migration to OS/2 and Windows/386 is being used to provide Sandboxing.. it’s all like Java never happened.

0
0
Bronze badge

back to the future with segment registers

More than most people realise. Segment registers on the 8086 segmented the memory space, just like on a real computer.

unix / java types don't recognise the analogy because their mini-computers didn't support a segmented memory, but main-frame operators know that isolating users, os, and applications was one of the purposes of a segmented memory architecture.

1
1
Silver badge
WTF?

Worshipping old Alchemy, I see.

Luckily these days we have paging.

And equating the clubfooted, useless 8086 segmentation with a "real computer"? Really, now.

1
0
Facepalm

Segment registers

@david12

No, segment registers in 8086 were just an excuse not to do supervisor mode correctly. The 68000 did it right but got to market 6 months later, which is why the world is 8086 based and not 68000 based. Indirectly that led to the ascendency of DOS and thus Microsoft and Windows.

0
0
Facepalm

Realy good idea

I wonder why no one has tried this before?

Its not like the internet is still suffering from ActiveX.

3
0

Native Client runs across all processors

"Native Client that can run across all processors. Portable Native Client – or PNaCl, pronounced "pinnacle" — compiles C, C++, and other languages into the Low Level Virtual Machine (LLVM) bitcode format"

Isn't this what JAVA was more-or-less originally all about, makes me wonder why it never succeeded.

"A Java Virtual Machine (JVM) is a virtual machine capable of executing Java bytecode"

http://en.wikipedia.org/wiki/Java_Virtual_Machine

1
0
Silver badge

Sounds good to me...

Does it have to be internet focused?

Google doesn't really get advertising directly from this (though I suppose it could be a route for GDocs to take). I suspect it's there to hurt MS and (now) Oracle.

How about moving non-web apps to the browser? It may be cpu dependent but is it OS-dependent? How much fun would it be deploy a new app on a web server and let the local site-proxies take the code distribution strain? Devalue the windows ecosystem (no need for software distribution, reduced requirement for anti-virus, local data backup etc) and decouple the apps from the OS and you reduce the likelihood of bing becoming prominent.

As for java, well it doesn't look very good does it. Not as shiny as things rendered in a nice chrome browser! Java is slow and ugly and not the most stable of gui's. Chrome has tab/process separation, a pretty interface and is running code presumably close to x86 code.

0
2
Thumb Up

@Christopher Key

PNaCl - pinch of salt.

Very subtle. Like it. You must write cyptic crossword clues for a living :-)

0
0
This topic is closed for new posts.