Is there still room for another OS to hit it big?

Post Reply New Topic

I still find Ubuntu annoying, and it tends to become user-difficult (if feature rich) when you go anywhere off the beaten path. That's a sign that the UI was either rushed, not done by the most talented people, or both. Depth is hard! I don't appreciate having to resort to Google to find out why I can't mount an NFS drive.

I can't speak for Mountain Lion, but I don't find Lion to be *that* polished - it still uses ridiculous .dmg files for installers, it's buggy, and Finder is a pain in the ass, to name a few. It's not like iOS, where you're hard pressed to find something that wasn't actually well thought out. It's a shame, because their concept of putting an Apple quality UI on top of a *NIX quality kernel was great.

Last edited by MyFutureSelfnMe on 03 May 2012, 1:21 pm, edited 1 time in total.

You just reminded me of a developer Q&A where Linux fans were trying to convince an Adobe software engineer to port Photoshop over, and his argument was something along these lines. Hopefully there are some angels in Red Hat's network willing to roll the dice.

My experience is in app programming, not OS design, so forgive me if the following thought sounds dumb: I wonder if it is feasible for VM apps to talk to the HAL of the host system? I remember using emulators that have at least depended on DirectX or OGL to delegate rendering to the best hardware possible.

VMWare has supported hardware accelerated 3D (and I think other functions) for a couple years; they probably would have faded into irrelevance if they hadn't. The overhead isn't tremendous as for most hardware functions you point the hardware to a memory buffer full of instructions/data/etc and tell it to go to town; the buffer just has to be mapped to the right virtual space. Some people have the misconception that an enormous, constant flow of API calls need to be individually translated. There is still overhead but the significance of that overhead is approaching zero as the years go on.

I believe all future OSes will virtualize and sandbox every single process running. Hopefully while also limiting the exposed API; if you think about it there is no reason for Microsoft Word to be able to do anything except draw to its little window and access the Documents folder, and the printer for example. With proper sandboxing, you wouldn't need to answer questions like "do you want to run this executable"; executable code could be part of web pages and could just run.

Would it really be as simple as making a bootloader-kernel hybrid? If you toss the specific features I mentioned earlier out the window, it seems like the GNU project might have the tools and the community necessary to make some strides in this area. The only issue I see down that line is the GPL having implications that scare away desperately needed attention.

I don't even know for sure that it would require a new OS; Google Native Client virtualizes 32-bit Intel executables entirely in user space (they based the technology on VX32; I had the same idea at the same time as Google but they had a team and I had a couple hours a day). IMO they never fully exploited Native Client. OTOH the last few graphics intensive applications I wrote were not anywhere near CPU bound; I think a 50% difference in CPU performance yielded something like a 10% difference in frame rate so this may all be moot in a couple of years and we might all be using JavaScript.

VX32 requires apps to be built specifically for it, it cannot just virtualize any native 32-bit code.

I'm not sure if user mode access to the virtualization features built into 64 bit processors is available on any OS.

Check out 9vx, it's a port of Plan 9 to VX32 that runs the entire Plan 9 OS as a plain user process on Linux/BSD/OSX. It starts up faster than OSX Text Edit.

Maybe not, but I imagine that one day demand will grow for a unified interface where applications programmed for different platforms could be executed without needing to switch between OSes or VMs manually. The rationale would probably hover around questions like "If I have 5 OSes on this machine, why would I want a separate GUI for each of them?" or "What if I don't want to be bothered learning how to best partition my drives for all these instances?"

If we lived in a world without patent or anti-trust suits, it seems that we would end up with an OS that can reduce competing platforms to mere run-time libraries.

I'm seeing split factions here.



I'll do that, thanks!

Sure. GNU Hurd... In 2088

HAAAAA.... That's good.

That soon?

If we ever get quantum computers there will be a new line open for operating systems.

ruveyn

Massively parallel computing a-la GPU shaders are going to become pretty commonplace. We're still going to be using the same types of OSes we already do.

As far as I'm concerned, android will begin to take over apple as people realize apple is treating them like a fish on a hook. Apple will bury itself by its business practices/inflexibility, and won't see its profits grow for many more years. Linux pc's might become more popular but will probably never fully take off - yet more people will probably move over from windows as the younger generations become more tech savvy/and android devices become more functional to use in schools. I also think Windows will become a poor choice because of its inability to be stable/the amount of issues people experience, so they will also start to lose ground. Blackberry seems to be recovering slowly from its own shortcomings, but is still a good choice for business so they might stick around..... if they make themselves appealing rather than bothersome in the future.

Android already has a pretty good user base and is becoming more flexible/still remains fixable through patches/and or editing of the OS by users. Plus with their chrome books and stuff they aren't gaining much ground but are demonstrating an idea.

Another OS of any sort on the market isn't really viable - when you look at what HP did - everyone just figured out a way to root them and put android on them. But, I'd foresee popular OS's continuing to develop for specialty items....

But it really makes sense to use an already flexible operating system to control other new technology's. They may go by different names, but with the framework there to create anything, and everyones money/time put into each individual OS - the companies sort of ensure they will continue to have a user base/competition will be null. Since people are more concerned about cloud storage/entertainment/easy setup nowadays on demand.

I disagree with so many of your points it's hard to know where to begin:

- Android will take over Apple when it provides a superior experience, unless Apple becomes too expensive, which they won't allow it to be because they choose their price point intelligently. Apple is only too inflexible for developers and serious power users, and those represent single digit %. The likelihood of Android ever providing a superior user experience to iOS is somewhere between 0 and none. iOS has already mostly nailed it and they have very little to polish.

- Windows 7 is clearly more stable than Ubuntu 11.04, which is the last desktop Linux I used. Linux is still ahead in a number of areas like disk performance, but there is no possibility Linux will take over the desktop. Mac OS X is the only other serious alternative to Windows, and it's pretty flaky too. Unlike iOS, Apple failed to nail the user experience on OSX.

- RIM missed the train and it left the station sometime around 2005. If they built this: http://crackberry.com/blackberry-blade- ... ot-be-true as a real phone, they could turn it around, but they're too meek and too Canadian to pull something like that off at this late hour. I hope they prove me wrong. Their existing phones have felt like they were developed by people who didn't want to be working on it, for a long time. There will be no slow recovery for RIM, they will either make a comeback like Lazarus or they'll fail and their assets will be bought out. If the company didn't already have such a good debt ratio they would already be toast.

- The kind of people who use Chrome books and modify Android are as mentioned in single digit %. You are right though that if a company like HP tries to pull a fast one and build their own OS, they can't compete.

- It pisses me off that Android uses Java.

Cheers

Last edited by MyFutureSelfnMe on 05 May 2012, 3:04 pm, edited 1 time in total.

AS-OS ... an OS that get things done efficiently, with minimal distractions, no clunky GUI , interfaces well enough with other OS's to get things done but not well enough to catch their trojan viruses (its too busy with its next task to even notice them).

Sounds like a winner. Go for it.

Ios will fail under its own success. There isn't much to their devices but sleek looks and cheap software gimmicks. Android devices are catching up in hardware, past that it'll be a fair game.