Linked by Thom Holwerda on Tue 25th May 2010 21:10 UTC, submitted by asupcb
OSNews, Generic OSes EyeOS has released version 2.0 Beta. "After several months of hard work we're happy to announce the immediate availability of the official release of eyeOS 2.0 Beta. And even more: the new release doesn't come alone but with the brand new website, which has not ben redesigned for the last 2 years now. eyeOS 2.0 Beta can be downloaded from the new downloads page and tested from a Beta test server in"
Thread beginning with comment 426962
To view parent comment, click here.
To read all comments associated with this story, please click here.
wannabe geek
Member since:

I know, CS people always complain about the abuse of the "operating system" concept, but I think the strictest definition (as seen in Tanenbaum books, for instance), which basically restricts the operating system to the kernel, is not particularly useful these days. For instance, there are microkernels, nanokernels, exokernels, hypervisors and whatnot, which signal a clear trend towards getting as many things as possible outside of the kernel, and I've even read discussions about what a kernel really is.

But if you stop and think why you should care at all about what operating system you are using, it turns that you care because developers write applications on top of it. If your operating system has no applications, it's useless. So, let's suppose that most developers decide to just write Firefox plugins; then Firefox (particularly its API) would be, for all practical purposes, the OS. But, you may object, Firefox needs to run on top of, say, GNU/Linux. Fair enough, then the OS is Firefox+GNU/Linux, but the main layer is still Firefox, because it's where the action is, where all user applications are built. You could replace the underlying GNU/Linux with, say, FreeBSD, Haiku or your pet kernel that only supports your own hardware and can only run Firefox, and then you would get all the applications for free.

I picked the Firefox plugin example precisely because it's NOT really what is happening now.

Instead, I think a good summary is, in Gilad Bracha's words, "Javascript is the assembly language of the internet platform (and the browser is the OS)". I would add that W3C's standards like HTML, and WaSP tests like Acid3 are then the rough equivalent of POSIX.

But it's not about how you name it, be it an OS, a desktop, a platform or whatever. The main point is that the "OS battles" are about what developers will take as a basis to build ever more complex applications on top, not about the details of hardware drivers which are functionally identical from the user's point of view, and which, by their nature, can only have a limited complexity.

Reply Parent Score: 2

bogomipz Member since:

I have to disagree. It *is* important how we name different concepts because otherwise everything gets confusing quickly.

The appropriate term is "software platform" or just "platform". The latter is already getting confusing because we also have "hardware platform", but you can understand which one was meant from context so it's usually ok to use the shorter form.

Firefox is a platform, eyeOS is a platform, Ruby on Rails is a platform, KDE is a platform, Emacs is a platform, and even Vim is a platform.

None of them are operating systems because they don't control the hardware. These are all layers on top of the operating system to provide a high level platform for people to develop software on.

Reply Parent Score: 2

wannabe geek Member since:

Well, I agree it is important to call different things differently, and also to use the same name when two things are essentially the same. What I meant is that it doesn't really matter *which* name you use (or which name is adopted by convention) as long as you are clear and consistent about it.

I see a lot of arguments about whether something is or is not an OS, as if it made some crucial practical difference. But the textbook definition is increasingly irrelevant, I think, for the reasons I outlined. So, you can either keep that textbook definition and start to talk more often about "platforms" rather than "operating systems", or you can change the definition of "OS" so that you can go on using the term.

Back to the concept of the OS as what "controls the hardware", we should remember that it doesn't do that directly; it only tells the CPU how to do it. You may also see a blank computer as having such a crude operating system that you can only run one program at a time, and you have to reboot to switch to another program. This program then would be the "platform", which manages the resources given by the underlying OS. In a similar way, KDE is actually controlling the hardware, by telling the underlying layer what to do.

Another thought experiment (or maybe not so much?); suppose you bake Linux and everything except the web browser into the hardware. Now, of course, the OS of this machine is the browser. But it's the same browser you just dismissed as unworthy of that name.

The OS is not the only program which controls the hardware, it's just the first layer of software on top of the hardware. But different people understand this "first layer" concept differently. For most people, the software that comes with the machine when you buy it is the first layer, but many textbooks focus on how "low level" the code involved is. I think that's a pretty useless definition, because the boundary between software and hardware can move one way or the other with each generation of hardware, but the underlying architectural complexities are the same.

By the way, this reminds me of all the brouhaha about the changing definition of "planet". Pluto is exactly as small and faraway as it was before. What I mean is, come one, it's just a word. My contention is that the concept of OS is almost as blurry and pointless, and just as emotionally charged (for some), as that of "planet".

Reply Parent Score: 2