Submitted by With all the recent hype surrounding Google’s Chrome, it’s refreshing to see someone taking a few steps back and looking at the bigger picture. Superlatives were abound about Chrome (I personally really like it), but some people really took it overboard – take TechCrunch for instance: “Chrome is nothing less than a full on desktop operating system that will compete head on with Windows.” Seeing my nationality, I know a tulip mania when I see one. So does Ted Dziuba.
All it takes for him to explain that no, a web browser is not an operating system, is three images. The three high-quality images, made in a simple image editor, pretty much sum up why I believe this mythical web operating system, over which every web journalist wets its pants, will remain a pipe dream. Dzubia sums up:
Users have pretty basic needs when it comes to computers. They want word processing, spreadsheets, communications, and games. These needs have not changed much since the advent of the personal computer. So, when your Aunt asks why her 1.2GHz computer isn’t fast enough to run an online word processor that has the same fucking features as the 1987 version of Corel WordPerfect, you don’t have an answer for her. There is no justification.
Google Chrome is a nice web browser for a beta release, but it’s not really revolutionary or all that earth-shattering as some make it out to be, and it most certainly isn’t an operating system. Ground, feet, keep. You put them together.
Surely you’ve seen the new Crysis javascript demo…
Ted Dziuba also wrote this on the same subject: http://www.theregister.co.uk/2008/09/08/dziuba_chrome/
Check out youos.com… errrr never mind they shut it down. It was an Java script OS run in a browser.
Krreagan
I’m not surprised it has been closed. I don’t see any compelling reason to use it. What about even one more layer and developping a so-called “operating system” that would run inside WebOS (or YouOS, if you want)?
An operating system is a computer program that is used as an interface between hardware and software applications so that an application can store information onto a HDD, communicate with a sound board, receive input from a keyboard, etc… YouOS is not developed in ASM or C AFAIK. We could call it a web-based personal manager maybe? Nothing more.
http://en.wikipedia.org/wiki/Operating_system
There’s still eyeOS:
http://eyeos.org
BTW, this concept of Web-OS is something you install in a server and access from any browser. The server is then the analog of your computer, where you install all your stuff. Chrome is more like a Cloud-OS, since there’s no unique place where your programs and data are. It’s more appropriate for SAAS.
Some are objecting to the use of the word “Operating system” for this kind of platforms. That’s a matter of terminology. Strictly speaking they are probably correct, but what counts is what the platform is, the thing that application developers target. If Chrome becomes the platform, then the OS becomes like a driver, just and “implementation detail”, as Sebastian Kuegler from KDE put it.
What would be desirable, then, is to cut off a few layers of the stack. Decide what the desired platform and behavior is, and then get it as close to the metal as possible. An example of that is the 9p protocol from Plan9. I mean, if we want computers to be wired to each other, why are their communication protocols so verbous?
how do you explain the same aunt that the same computer cant run notepad?
that google writer isnt feature rich is not important, as long as it has the features that the silent majority use…
Note how the final ‘stack’ ends up looking like a giant pile of dog sh*t…
Note how the final ‘stack’ ends up looking like a giant pile of dog sh*t…
If I had that many layers running on my machine it would indeed atleast run as fast and be just about as pretty a sight, too..
I tried out Chrome on Windows XP. It clearly is a browser with tie-ins for Google web applications. One thing that makes it unique is the ability to determine CPU usage per tab. I kind of liked this option. I could see a tab running at 90% with a flash banner ad and killed it. I wished for even finer control over Flash, such as the ability to enable/disable by a filter similar to Firefox extensions available. Compared to IE 7 it ran Google apps faster showing the benefit of Chrome’s speedy JavaScript engine. I can see how people are referring to it as a ‘web OS.’ It appears to be moving towards a web application container for running Google-centric applications.
Edited 2008-09-08 23:33 UTC
It is not just goggle centric apps. It will run any application just fine.
An OS directly deals wih the hardware as the host of your applications. With Chrome, and all other Web browsers, you need an OS to host that! Even if the theory of a WebOS was true, would you entrust your Hardware to the Web? Imagine the exploitation possibilities!
stop talking sense, man! That has no place in the noughties!
I agree with you about the “There’s No such things as a WebOS.” If you maximize the Chrome browser window and your Google web apps persist their state on your local PC you essentially have a desktop environment with a fancy web interface. You could call it a WDE (Web Desktop environment). This is a layer on top of the applications layer. It is a pretty common error for new Linux users to call their environment an OS when they are really referring to KDE, Gnome, XFCE, etc.
I agree more with your statement in regards to Desktop Environments. But even the idea of an online operating environment will not fly far simply because of privacy and security issues. Basically and good legal team of said service would deem everything that you create and store on their servers as THEIR property or give them exclusive rights to the created or stored data. If you doubt me, just read the EULAs or their Terms for use. This will not fly in the corporate world, nor far with your average joe on the internet. People already store their documents in their emails anyway. Not to mention flash drives, etc.
From the blog:
“After a while, everybody wanted to be a programmer. Since programming is actually kind of hard, many of these folk landed in PHP and HTML, hence the explosion of webapps.”
I don’t think the kind of ‘everyone’ who landed in PHP and HTML are the same people responsible for web apps – ‘everyone’ might write some javascript or basic PHP site, but web apps? The likes of Google and Yahoo are hardly ‘everyone’.
Also, it’s hardly fair to compare Windows ‘stacks’ and Linux/Windows virtual machine stacks. Remove the virtualisation and you have a stack no higher than the Java example. I could make the same argument about your favourite Linux-only subsystem, with Windows & VirtualPC on the bottom. If you don’t like the size of your particular linux/virtualisation stack, the problem is not Chrome itself but that it doesn’t run natively on your platform. Which is something that will surely be addressed in time – Chrome’s only days old so far!
Edited 2008-09-09 01:33 UTC
The final stack is an attempt at misguiding at best. Like you say, you could argue that the browser takes the place of the Language Runtime and you end up with a similar stack as shown in the 2nd picture.
Though to be fair, the browser will have to invoke some sort of VM which would add an extra layer to the 2nd picture. Still, that’s not as bad as the article tries to portray it.
As does Ted. The point isnt running Chrome on top of Windows on top of Linux. Thats just pure stupidity and he knows that. Imagine running Chrome on Asus’s Linux based Bios. Instant on machine. Fully web ready and runs any application your ISP provides. All of a sudden you can have your cable modem do 90% of what your $2,000 computer can do. And guess what? It doesnt get viruses or malware. And even if it did, Your ISP could re-image the thing in about 5 minutes. No more draggin your machine to Best Buy or Circuit City to get Geek Squad twice a year. Will it replace computers for everyone? Of course not. But 75% of people with computers are still computer illiterate. This could go a long ways towards fixing the PC as an always broken appliance.
Well.. you said it yourself, “Linux based bios”.. First of all, the bios is not Linux based, the bios is the same as its always been. They just decided to have a copy of Linux on some solid state memory on the motherboard somewhere. It is still an operating system, that the web browser is running on top of.
It doesn’t matter how fast the operating system loads, that does not mean it isn’t there.
As a general rule of thumb, if a piece of software requires an operating system to run, there is a very very very good chance (about a 1000000000% chance) that that software is NOT an operating system, nor will it ever be.
Now repeat after me. “A web browser is not an operating system”. Again.. “A web browser is not an operating system”. Now, go find a chalk board and write that 500 times until you understand.
And why in the name of god would you want the entire content of your computer to be stored on a server at your ISP?? Are you out of your mind? People talk about how horrible closed source software is because you don’t know what its doing and you dont have control over it, and then they turn around and talk about how good an idea a thin client with files stored at the isp are????? Really?!? Talk about giving up any and all control of your computer system. In that situation Sony wouldn’t have to install root kits on your machines, they would just lobby the government to force isps to install them on their “web computers”, and you wouldn’t have a damn thing to say about it. Not that they’d have to, they would just have the ISP remove any files they don’t like from your system.
“No more dragging your system to best buy”..
This is exactly what he is talking about. Sacrificing performance for everyone, to accommodate those who don’t know what they are doing, but still want to pretend they do. A lot of people never take their systems anywhere to have it fixed, they know how to copy files, put a CD in the drive, push a button, and push the next button when they are told to. And those that do, is it really that hard to take a 20 minute drive to a best buy, say “fix this”, then go home and watch TV for a couple days?
P.S.
As for “dragging”.. come on, they aren’t THAT heavy. And I have a really large case..
Edited 2008-09-09 05:03 UTC
I completely agree that a web browser (i.e. google chorme) is NOT an operating system. But that doesn’t mean an OS has to run on bare metal, it can be hosted in another one, have a look at Inferno by VitaNuova for a good example.
First I never said that the browser was an OS. But that doesn’t mean it won’t be a core part of the operating system allowing web apps to be run with greater efficiency than today. Second, I already concluded that this would NOT replace all computers. This will end up being (I think) technology that finally allows thin client like appliances. It has a long way to go before then. Last, who said that all the data would reside at the ISP? How easy would it be for your ISP appliance to have an esata port on it and you just hook up an external drive? The point of the device being light weight and managed by the ISP is that it reduces the current problem of viruses and malware by not making the end user the administrator any more.
I don’t trust my isp with the power they have already.
Sure as hell i won’t want them to have the abilitly to ‘re-image’ lol.
It’s my damn computer and i should have the controll over it.
The thing you describe is not usefull for the user.
It’s just like digital tv. Only usefull for the providers/compagnies, so they have more controll over you.
yeah, but we’re not talking about you. I assume you are computer literate, being on this forum and protesting losing control of your computer. Most people aren’t computer literate. There fore they have no control over the system and as such, can’t really lose control to their ISP. Further more, as has been seen in recent years, these computers that the owners can’t control end up being bots in a botnet. That’s bad for everyone. These are exactly the type of people who could benefit from this type of service.
If they are illiterate, then teach them.
They are better of making mistakes, and learning from them then keeping them ‘illiterate’ and worse making the next generation even more illiterate by making them stop thinking all together.
All it takes for him to explain that no, a web browser is not an operating system,
The analogy with the idea that Java makes life simpler, but with a slight overhead that we do not mind because of our multi-gigahertz processor, is false.
There are some optimizations that cannot be done at compile time with a C program, but that a Java Virtual Machine can detect and perform at run time. So it is not because there is another layer that automatically it will be slower. Sometimes it can be faster.
If the majority of Java coders does not care about unuseful object creations, about releasing some object references or forgets that the garbage collector is not magic, they should first learn to code in C.
Actually, only C programmers should be allowed to code in Java.
I advise you to have a look at profile guided optimizations (PGO) which are supported by GCC, Intel, MSVC compilers.
With PGO, you build a binary with profiling enabled and you run it through a typical use case scenario. The profiling data that is generated will then be used when you compile the app for the 2nd time, allowing the compiler to determine which code paths are more frequently used (i.e. Hotspots) and more aggressive optimizations can be performed (e.g. function devirtualization, deeper inlining, etc).
The end result is, you get the benefit of optimizations that could only be done at runtime. The major difference between PGO of statically compiled languages and VM run languages, is that the profiling is done during development and not when your poor enduser is running the application. It’s easy to see which one is the superior approach.
Ted is clearly not seeing where Google or the future of web apps are going. Look at what happened with email readers. They used to be run on your local computer from vendors like Qualcomm (Eudora) but nowadays people are more and more running their mail online.
Google isn’t targeting Joe Average or Ted’s granma in their first Chrome release. Google have a clear strategy and part of it is to offline enable their web apps making them more like regular apps. Some of the technology in the Chrome browser let’s them do just that.
Where Google is heading is obvious, the question is how long it will take before it’s gone through early adopters and are being picked up by Ted’s grandma. Here’s an insightful blog in the same matter http://code.qbranch.se/post/show/33
Most people in this forum seem to care about efficiency, slim design, “sense”, whatever. Obviously, there are many developers, coders and power users here.
Compare that to the general “Aunt Mary” or “John Doe” … neither of them cares how complex a piece of software is. It’s a lot of more complex than they can understand, anyways. They care about ease of use, availability, perhaps price.
On the other side, there’s this big company. They want market share and are willing to give a lot for it. Traditional applications don’t fit their business model, as those tend to keep all data local to the computer’s hard disk. No urgent need to “go online” means no chance to intercept for advertisement.
So they put their full weight into getting people using web apps. Web apps mean chances for advertisement, chances for data indexing, chances for “targeting”, whatever that means.
Everybody who tried to make and sell something already knows how difficult it is to get the message out to the customer. Average people don’t seek for the good source code design, but for a nice look. Often they use whatever they find first. Very often they take what they saw in some ads. It’s the marketing company reaching those people and selling items, not the developer making headaches about designing his code.
So, traditional applications do have a lot of competition here. Wether chrome is a reasonably designed browser, a virtual machine or a complete OS on top of another OS, 95% of the computer users just don’t care.
Traumflug
Edited 2008-09-09 08:26 UTC
I just wish I could use one processor exclusively for PGO and GC during runtime, but I appear to be ‘protected’ from doing that.
Mad props to whoever gets my post title riff.
If you’re running a machine with more than one processor, the JVM and I’m sure the .NET CLR will will utilize the extra processor.
If you’re using Python, you’re out of luck.
No matter how you look at it Google is NOT and will NOT develop an OS for the foreseeable future. They have said that quite a few times now and I really don’t see what the big deal about a new OS is. Sure Windows sux in every one of it’s incarnations, Linux is hardly ready for prime time, and MacOS has it’s share of peeves but at the end of the day Google is a business and they really don’t care too much about the fact that MS Office does not run on your OS or that you have a problem with Skype. Every software ever released ( or bought ) by Google ties in directly with their business plan. Google gave a chance to both MS and Firefox to get their crap together and ready themselves for the services that Google is offering but neither one of the 2 products is even close to achieving the performance that Google needs/wants. Hence, the new browser. If you look just a couple of years back you will recall that Apple did much of the same with Safari for Windows. As to why Google is creating a brand new browser rather than building on top of Firefox … well Mozilla is in charge of Firefox and steering Mozilla in the *right* direction might prove quite a challenge even for Google. Plus you almost never rely on a 3rd party for something that is considered a mission critical resource for your business.
So don’t get your panties in a twist and stop with the sensationalism.
The C64 was a machine without an OS, very common for 8bit microcomputers. It could only run one application at a time. But there where competitors that did have an OS like the Mac. The Mac could kinda run more than one program at a time (ok you could do the same with the C64 but cut n paste and reboot and load “*”,8,1 would be a pain). So they introduced GEOS to the C64 and lo and behold it could run more than one app at a time.
Now imagine that you have a souped up C64 that runs google chrome. Only one application at a time. But chrome can run applications inside of itself. So by applying chrome to your souped up C64 you have a Operating System. Anyone who doesn’t understand how you can get a computer to run a single application without an OS is thick. Sure that does not mean chrome is an OS, but it does mean that it can be one.
I’m not familiar with the C64, but did every application talk directly to all the hardware it used? If so, then there would be a lot of redundancy between applications, since they would be doing the same things. If not, then there was an OS,even if it wasn’t visible to the end user…
There is a Kernal in the C64 but it is more like a library than an OS.
Why is this modded down? It’s very true. The KERNAL had one large jump table into it, and very little of the code was actually hardware I/O (mostly memory-mapped). Almost all the KERNAL could be switched out or shadowed in RAM, allowing you to modify or remove the KERNAL.
The C64, and every other computer have operating systems.. In fact, repeat after me OPERATING SYSTEM….
The system, whereby the machine operates…
In fact, if you would have searched you would have found this..
KERNAL
From Wikipedia, the free encyclopedia
Jump to: navigation, search
This article is about Commodore’s 8-bit OS software. Kernal is also a common misspelling of kernel.
The KERNAL is Commodore’s name for the ROM-resident operating system core in its 8-bit home computers; from the original PET of 1977, via the extended, but strongly related, versions used in its successors; the VIC-20, Commodore 64, Plus/4, C16, and C128. The Commodore 8-bit machines’ KERNAL consisted of the low-level, close-to-the-hardware, OS routines (in contrast to the BASIC interpreter routines, also located in ROM), and was user callable via a jump table whose central (oldest) part, for reasons of backwards compatibility, remained largely identical throughout the whole 8-bit series. The KERNAL ROM occupies the last 8 KiB of the 8-bit CPU’s 64 KiB address space ($E000-$FFFF).
The KERNAL was initially written for the Commodore PET by John Feagans, who introduced the idea of separating the BASIC routines from the operating system. It was further developed by several people, notably Robert Russell added many of the features for the VIC-20 and the C64.
I think you are confusing an Operating System and a GUI, as in Geos.
Don’t be so dense it is possible to make a computer that has no OS. What came first the computer or the OS?
No I mentioned GEOS not because it can draw to the screen but because it has one thing that many early OS’s had the ability to switch running programs. An OS is the layer between Applications and the hardware not an application and the hardware. If you only have one application running ever you don’t need an OS. What you may need is a set of libraries for similar applications but those libraries are not OS’s.
You sir, are dopey.
Of course the computer came first, but without the OS, it would sit there doing nothing.
No I mentioned GEOS not because it can draw to the screen but because it has one thing that many early OS’s had the ability to switch running programs. An OS is the layer between Applications and the hardware not an application and the hardware. If you only have one application running ever you don’t need an OS. What you may need is a set of libraries for similar applications but those libraries are not OS’s.
CP/M, MS-DOS and all those are still OSes even though they don’t boast multi-tasking capabilities. An OSes task is to allow the user load applications and manage their hardware. It also provides the applications with a common set of features they all can use without all the applications having to develop their own versions of hardware layers, memory managers and all. That is the task of an OS.
In short: OS is something that initializes everything to a common and known state and provides applications with method for utilizing a pre-defined set of functions like opening and saving of files, handling filesystems, memory and I/O. It is still an OS if it doesn’t support multi-tasking.
It’s one of those cases where the writer managed to miss the target by a mile, and somehow hit it.
First, his A => B => GREEN diagrams explain nothing. There isn’t exactly any actual correlation between the move from C to Java, to a WebApp. First, Chrome’s HTML renderer and Javascript engine are using the browser as a VM, theoretically putting a chrome-based webapp on the same level as the Java/Perl example.
Second, and this is plain obvious, using a Windows emulated environment as a reason for “Chrome is not a browser” is counter intuitive, and misleading, since, on Windows, Chrome runs “natively”. He could’ve tried running it in an emulated environment running on Cygwin, and claim this meant something.
However, he is right on one point. A browser is not an OS anymore than a webapp has anything to do with computer software, running natively, emulated or on a VM, and that has to do with one major element, which is the actual Internet.
The Internet runs on a series of protocols that, apart from other similarities, share one specific idea, which is that communication is based on a simple idea: You send a request, you get a reply, end of communication. You can split your “application” to many small parts and have each of them send requests and process replies simultaneously and asynchronously, but it’s still the same request-reply process. There isn’t anything there that can even come close to emulate or imitate a non-web application’s richness of abilities and transactions. The way information flows between the front and back ends of a non-web app leaves the web-based apps way behind.
Apart from that, there’s the question of what a web-app is built with. While back-end technologies have become amazingly rich and powerful, the browser, at its essence, understands two things: HTML and Javascript. Whatever you cook on your server have to ultimately be translated to those two components. While plug-in techniques (Flash, AIR, Silverlight), have improved massively in recent years, those have no need for the browser as a platform, and definitely don’t have anything to do with the Web-as-an-OS vision that some people hold.
On top of that, the idea that browsers, for security reasons, have little to no actual access to system resources, can’t change or create system elements and can only use whatever freedom the browser allows them, makes the dream of the “WebOS” remains a dream, and not a very imaginative one.
UPDATE: If you really want some cool picture, this page might do the trick: http://www.theregister.co.uk/2008/07/07/web20_for_developers/page2….
Edited 2008-09-09 12:03 UTC
I think the big application of Chrome is when it goes mobile. Imagine having your Nokia internet tablet or something smaller being capable to edit all your office files & full access to your Gmail in a package smaller than these netbooks/sub-notebooks. Instead of having the power and storage space on the mobile device, it can be housed on the google servers.
It wouldn’t be for everyone, and it wouldn’t replace the current desktop model. I could go into greater detail, but I am at work. You guys are smart, you can see how it would play out.
Robert Cringely recently wrote an article (http://tinyurl.com/5heozg) that talked a little about his ideas about Chrome. I think I put it best in my comment to his post which I’ve copied below (unfortunately I can’t link to it directly or I would have). That is not to say that I agree with the idea of Chrome as an “OS”, just that viewing it that way answers more questions than not b/c from Google’s perspective the Internet is the OS, and the web-browser is the way to access it. With that in mind…
Haven’t used Chrome (yet), but I highly doubt Microsoft would be able to get away with turning off Ads in IE. Why? B/c it is a recognized monopolist, and doing so would harm a big portion of the market, not just Google. They do it for even a day, and they’ll be paying Google and everyone else the money lost. So, no – that’s not Google’s fear.
However, the “OS” scenario is a lot more likely. What is it really that Chrome gives them? (Aside from the irony that Microsoft couldn’t release IE “Chrome” due to the Antitrust trial…a tech that later became WildTangent.) Chrome gives Google control over the browser to a new level. They can write their apps to their platform. If a browser doesn’t support what they want to do, then they can simply point the user to their own browser, and say “Hey, use this if you want to do X.”. They can also ensure their applications work on any platform (Win32, Win64, Solaris, Mac, Linux, Unix, etc; on any processor) they choose.
Yes, Chrome may help them protect their “Ad” profits; but it really gets them more than Ads will buy them. After all, they also have their Google Appliances that companies purchase and rent, and the Google Productivity suite a lot of companies use directly from the web, and more. It let’s them extend their software platform against Microsoft and do what they want, and do it all on the web.
So yes, the “OS” answer is likely the more correct one. After all, the Open Source browsers may not fully implement what they want either…now they can control, or at least provide input and pressure to support the features they want, their base platform – the Web.
People cannot agree on what an OS is. To me, and to many, it’s the software that drives the hardware full stop. Everything else is on top. So, the Linux kernel is an OS and a distribution is the OS + some. But for some people the OS is the collection of the kernel and the standard operating environment, like the BSD base system or even all of Win32 for NT. This is why you get arguments between people falsely comparing Linux to Windows when they mean KDE to Explorer.
The “web OS” people suffer from this same terminology problem, made worse by the fact that they (through ignorance or some misguided attempt to be radical) talk about things which might be called an operating environment as an operating system. People will accuse me of splitting hairs and semantics, but precise terminology matters when having technical discussions.
Let’s break it down. The way I see it you can summarize the distinctions as four things
– The OS
– The Platform
– The UI
– The Environment
some of which can be said to overlap in many areas. What a user sees every day is the UI. What he works in is the environment. What developers are concerned with is the platform, but a little UI and environment is thrown in. What the machine cares about is the OS.
What these “web-top operating systems” or “web desktops” as I have sometimes seen them called are striving for is to be a platform and an environment. They never have and never will touch the OS.
Some cloud computing enthusiasts see a day where the local OS doesn’t matter. Where only one application exists–the browser–and everything is run through it, or in/on it, or whatever. They see this as a kind of utopia, but while it might be radical as an idea or even useful as an abstraction layer I don’t think it will never happen to that full extent. These enthusiasts are the ones who willfully talk about a “Web OS” and do a disservice to those who don’t know enough to decipher what they mean.
The browser may–will, in my opinion–eventually be the major platform for application developers. The UI will be written with HTML, CSS and their successors, Javascript will drive the front end. Some code will be on remote servers and URLs will be called like both like libraries to deliver just what the author wants at the moment and as an integral part of the application. But the browser and all apps within it will never–can never–replace the operating system.
Hi…
That was an excellent summary. I have a question though. The caveat being that I am a mere commonplace user of computers!
Going by what you say in your post, does it mean that it is impossible to conceive and devise a computer that runs without an OS?
If the answer to this is no, then what would such an architecture look like?
And, if it is not possible then why not – what are the technological bottlenecks? Do you expect such bottlenecks to be perpetual – meaning that will they never be/ can they never be resolved?
Thanks and apologies again for what may be a dumb question.
Edited 2008-09-09 16:53 UTC
Consoles are a good example of a computer that runs without an operating system. I think the latest generation have operating systems (I don’t own one) but if you look at the older systems like PS2 and cartridge machines they had 2 parts, a BIOS and a game, there was no OS to speak of.
You mean like thin clients…?
No. I’m talking about game consoles. Nintendo, Playstation, Sega Genesis, etc. They bootstrap directly into the application. Sometimes there are common libraries used by the games and routines provided by the manufacturer but since they aren’t common across all games and each game ships with its own possibly different version they can be considered part of the application, not an OS.
Hmmm…so theoretically then it is possible to have computers without an OS. So, how practical would it be to have computers without an OS to do the kind of mundane tasks that we do everyday?
In that case each game implements its own customized operating system.
If I take the Linux kernel and port it to my cell phone, then customize it to do just what I need, then add all of my phone functions as kernel modules, what do I have? Is it an application or an OS?
Same thing.
Hmmm…OK. I see I am wrong. It is indeed impossible for the computer to function without an OS. Now the question is whether this is only possible paradigm for conceiving computers. Have there been any attempts (albeit failed attempts) to design non-OS dependent computers? Or, is that a contradiction in terms in the sense that a computer by default implies the presence of an OS?
Thanks in advance for bearing with these possibly naive questions.
You cannot run any computer without an operating system. Anyone who says other wise is lying, misinformed or playing loose with his terminology.
Another poster gave video game consoles as an example of hardware with no OS, but this is wrong. The operating system is what drives the hardware. If the hardware does something it has an operating system. It may not, perhaps, meet a standard definition of an abstraction between userspace processes and the hardware, but if it drives your CPU it’s an OS.
This is why you cannot–by definition!–do without an OS. Something has to make the hardware do things and that something is the OS, no matter what else someone may want to call it in the future.
I loved the link to Ted Dziuba’s article. The style of his diagrams matched closely the style of his writing. Something about “Glibc and Shit” struck me as very funny.
Oddly enough, his final diagram was exactly the method I used to test Chrome. I ran Windows in Virtualbox, and installed Chrome. As Dziuba’s diagram shows, it’s an ungainly stack. It’s all the more surprising that it worked.
Chrome ran fine, and at speed. Then I compared it to Firefox 3.0.1 on the same virtual machine, and both ran fine. That said, Firefox was faster. I run a tightly controlled browser, and Firefox’s overcame it’s rendering deficiencies by:
a) not loading third party adds, and
b) blocking most Javascript.
Chrome mostly closed the gap when I loaded a hosts file from mvps.org. That said, firefox does a better job blocking content that Chrome. A case in point: OSAlert loads third party content from com.com, woopra.com, and google-analytics.com. That’s three mostly tracking based sites, sites that require three separate DNS lookups, and three potential sources of Web site slowness. That’s not even mentioning the drag caused by third party advertising services.
In other words, the hype over Chrome’s alleged speed masks the real problem. Browsers are slow, because Web sites are slow. As Web sites rely increasingly on third party content, the potential for slowdowns increases, and that’s not something a new browser can fix, neither (unfortunately) can the Web sites themselves.