Slashdot Mirror
Parallels Beta Adds Boot Camp, Desktop
Verunks writes "Parallels has released a new beta of its virtualization product for Mac OS X. This new release includes one major new feature, something Parallels calls Coherency: "Shows Windows applications as if they were Mac ones. Try it and enjoy best of both worlds truly at the same time. No more switching between Windows to Mac OS." Check out this Screenshot"
More interesting to me is the Boot Camp support so you can have a single partition to run IE7 in Parallels to test compatibility of a website but reboot to play video games that need a little more juice.
I've been wondering why a Linux distro doesn't do this automagically with WINE. It seems like such an obvious feature to implement, and would be great for people new to Linux or even those whose who don't know how to use it if it just ran as if native...
The constant improvement that this product has seen in its short existence is astounding. When you consider that it costs only $80 and has no competition at this time, it almost seems like they're working too hard on it.
If Parallels was publicly traded, I'd be buying up a lot of their stock. These features are too damned useful for Apple to not add to OS X at some point, and the best way would be for them to just whip out the checkbook and buy the company.
Actually, with modern multi-core processors and oodles of RAM, virtualization kicks pretty much ass. When I run parallels in fullscreen mode on my macbook, you pretty much can't tell it's virtualized. It's more responsive than the dell desktop sitting in my office at work. The only thing you really notice is that the video card doesn't support hardware acceleration, so stuff like games suck. Then again, the video card in my macbook is pretty crappy, so even with 3d support they would suck =/
MacBooks and MacBookPro's do support right mouse buttons. Tap one finger on the touch pad for left click, tap two fingers for right click (and drag two fingers around the trackpad for scrolling, or zooming with Control pressed).
I installed this as soon as it came out, as did many other Mac users. My Mac (mini DP Intel 1.67GHz, 2GB RAM) slowed to a crawl as soon as I launched it. I had to yank the power cable. I uninstalled it and all was well. This is a common experience. If you're just going to try out a new version, cool, go for it, maybe it'll go well. But please understand that it's a beta -- don't plan on getting any work done with this.
What do you call "non-native support for Linux"?! Apple laptops run linux _as natively as it goes_ for ages and this doesn't exclude the Intel based machines. I even could setup a triple-boot on an Intel based Mac (vs. all the dual-boots I had in the past). All running "natively" of course
Now, mod me down freely. My karma can't get any worse...
I've installed it and it is very similar to Classic on PPC macs under OS X. As with OS 9 apps on OS X, a full copy of the operating system is running, but the windows are drawn directly to the desktop (or at least appear to, with some glitching at the moment). I have the Windows task bar running down the left hand side of my screen so it doesn't get in the way of my dock (at the bottom) and desktop icons (to the right). Running Windows with the classic theme looks better as the shaped edges of Windows apps leave a little triangle of the Windows desktop which looks a bit poor. Lighten up the theme and it works quite nicely on the OS X desktop.
:-)
Apple really needs to buy Parallels or do something similar. It would make a huge difference to people moving from Windows to the Mac and eventually, Windows could go the same way as Classic MacOS has under OS X and just fade away. I don't think MS would be very pleased with this development though
"I have the attention span of a strobe lit goldfish, please get to the point quickly!"
It's so small, it doesn't get in the way at all. I used the kind of adhesive that doesn't leave residue when you pull it off and you can keep sticking it on over and over. I don't know, works for me.
The soylentnews experiment has been a dismal failure.
This is under development by major virtualization companies. VMWare supports it for windows as guest (in beta), and Parallels has said that it was under development for a future release. This is harder than it looks though, since you have to develop a full blown 3d driver for windows and Linux (used inside your virtualized environment) that will send the calls to the host operating systems, in the case of windows transferring DirectX calls to the OpenGL API. If you want to stay generic (to work on both hardware nvidia and ati), you have to limits the possibilities of the card, or else you'll have to make a driver for each type of card you want to support. That's the theory.
Of Code And Men
Emulation is slow. E.g. a PowerPC executing x86 code by emulation will be much slower than a native x86. There are tricks, like profiling the application and translating, rather than emulating the frequently used bits, but it in general there will always be a hefty penalty. And modern performance critical code will use multimedia instructions which don't have 1:1 mappings to a different instruction set.
But on an Intel Mac none of this is an issue, since the Windows app and a mac one run on exactly the same instruction set. Of course, the API the applications use will be completely different. Virtualisation is about running two kernels simultaneously on the same hardware. Now this is tricky, because OS kernels want to be in sole control of the hardware. The x86 isn't completely self virtualisable, i.e. you can't trap and emulate all the instructions you need to fool the kernel, so you go back to profiling and translating, at least for kernel mode code. Or you can trap many more instructions than you need to. But recent intel chips have a technology called VT which plugs the holes and allows self virtualisation.
So you can run the guest kernel code at full speed, and trap and emulate just enough to keep the guest OS under control of the hypervisor.
echo -e 'global _start\n _start:\n mov eax, 2\n int 80h\n jmp _start' > a.asm; nasm a.asm -f elf; ld a.o -o a;
I'm running this beta build right now - have been doing all day as I do the exciting task of catching up with my accounts (Quicken UK, Windows only). There's some graphical improvements to the interface - I like the better laid-out screen for picking the VM. There's still some interface no-nos (ok button on the left? Nope, shouldn't be the case on OS X) and I think the dock icon is trying just that bit too hard when it turns into a dancing egg timer as you save a machine's state, but overall things are better and things are fine.
I upgraded from a previous install, which means I had a disk image of Windows installed rather than a real partition. What I'm wondering is how Windows would cope with being booted for real on MacBook Pro hardware one moment, then booted again in Parallels another moment. Surely that would kick Windows activation into life?
Cheers,
Ian
It's driver dependent. Support for it is built right into OS X, the latest Boot Camp beta adds a trackpad driver so you can do it in Windows as well. As for Linux, I have no idea-- there are certainly no Apple-provided drivers.
~Philly
This is really good for Parallels and will be important for the company in several ways.
Obviously it is a big feature for users who might be interested in Boot Camp and Parallels. One license, keeping the same settings etc.
The thing that will bring the real benefits to Parallels though are related to development. Working with Boot Camp means that Parallels can access the Boot Camp drivers for Windows that Apple writes. Every time Apple updates their hardware they'll update Boot Camp with new drivers. This will make it much easier for Parallels to keep up with new hardware.
Boot Camp adds a driver for the touchpad that includes Apple's right click implementation. Suddenly it's in Parallels automagically. Apple ads a driver to operate the inbuilt iSight. Parallels can start using it too.
Shared documents are potentially great. Apple should work with Parallels to ensure things like the iTunes library (and iTS purchased music) is available in the Windows partition.
Apple have already said that they are not going to include virtualisation in Leopard because they are so happy with the performance of Parallels.
If necessary they'd buy Parallels to ensure that development keeps going on. They might do it anyway to reduce the costs.
the more they over-think the plumbing the easier it is to stop up the pipe
This is the reason I would offer: As soon as a competitor's product to the one you're offering is available natively on the Mac, you'll lose customers. While I agree virtualization is now offering acceptable performance for many Windows-only applications, this virtualization does not integrate well with a Mac user's workflow. Once a native version is available, users will switch. And as Macs gain mind-share and market-share, if this competitive product has cross-platform support, the prospects grow slimmer for single-platform applications. Now I realize the difficulties in coding for multiple platforms, however there are cross platform frameworks available today to assist with just such endeavors. And while it's likely significant effort, depending on the vendor, it may be a strategic decision to rewrite an app so it's cross platform. As a point of reference, I'm currently involved in a project which has two major suppliers of a particular function. One has slightly more bells and whistles, the other is cross platform with support for Windows, Linux, and OS X. Both have similar market share, but the client selected the platform with cross platform support over the platform with a bit more functionality. Just something to think about.
The Objective-C-to-Java bridge is being abandoned because it really didn't make things easier for Java developers and because it was a pain in the ass to write code with for everyone and to maintain for Apple. (However, RubyCocoa will ship with the next version of Mac OS X because it's a lesser pain in the ass on all accounts.)
You may know more people who have VPC or Parallels than not (I do too), but how sure are you that those people will be representative to the entire Mac market? To the market you want to aim your product at? (Unless it's "technologically competent user who has ever heard of Slashdot", fat chance.)
There's also psychology in it. At its core, the people that are now switching to Macs are not switching *because you can run Windows on it*. They are switching *because you can run Mac OS X on it*; the ability to run Windows on it just pushed them over the edge because Mac OS X doesn't have a 90%+ market share. If they were indifferent to what software they preferred, they'd be using a different brand of computers, and run Windows, not Mac OS X.
Most Mac users, even the ones propped up with VPC or Parallels (I plead guilty), ultimately want to run Mac-native software rather than Windows software. Parallels is life-support for existing software that people need to run, and even if it was free and shipped with all Macs and took up half the memory and disk space that it does today, it doesn't make Windows software into Mac software.
You don't need to think that Mac software is superior to Windows software to concede that Mac software has an advantage over Windows software running in a Mac simply because it gets access to all system APIs to things like address books and keychains and hardware support and preferences, and because it looks like everything else you run. Windows software just think it's running on an isolated box and won't become aware of the Mac OS X side of your computer unless you as a user go to some length and the software itself supports it, at which point the developer will already need to make way in their timeplan and budget for Mac-specific testing.
Still not convinced?
1. Mac market share is currently surging. More people, not fewer, will arrive at the Mac platform in the next few years, and building a dedicated version (and almost no well-designed application will need to be rewritten entirely from scratch) is becoming more and more economically feasible.
2. Would you want to bet your entire Mac user base on a competitor not releasing a native Mac version? Unless it's a turd, people will switch to that in a heartbeat. You will lose out months of sales as you rush a native product to market, or need to pull out of a market completely.
Aha! You like trackpoints because you keep your fingers on the home row.
Good for you. That helps explain your preference.
Still, you said not having three buttons was part of the problem without explaining why. And that makes me wonder since my experience shows it works quite well, even for applications that need 2 or 3 buttons.
I am also puzzled because you confirmed my complaint about 3-button laptops - that you have to move your thumb sideways for every mouse click - and then said Apple's interface is the awkward one. Have you tried them both long enough to get used to them?
Anyway, my point is that many people who are used to three-button laptops may well find Apple's one-button approach more than adequate. Having used both for thousands of hours, I have - to my own surprise - come to prefer the single button trackpad. Not only for native OS X applications, but for Windows (under parallels and via VNC) and X11, using 2- or 3-button emulation. So much so, that I prefer using my laptop to interface with my Windows box rather than the very nice keyboard and mouse on it.
Apple's trackpad is very good. I'm not asking you to like it. But I'd like you to accept that it will be adequate (even excellent) for many users.
Check out Oxygen, its a cross platform XML editor.
http://www.oxygenxml.com/