Rogue Amoeba - Under The Microscope

What Does Virtualization Really Do To The Mac World?

Thu, 03 Aug 2006

Here's Some Hot Air...

Everyone's all geared up for the Worldwide Developers Conference (WWDC) next week, including us here at Rogue Amoeba. Mike Ash will be representing the team out in California, so if you see a guy in a Rogue Amoeba shirt who looks like this, buy him a beer. That, or tell him I said "Get back to work!".

Anyhow, several developers are posting predictions/wild-ass guesses about what Apple will introduce or at least show off at WWDC. Brent had some interesting thoughts about the basic UI direction of OS X, most of which I find myself concurring on. Gus also posted some thoughts that are worth looking over.

I suppose I might as well share my insider knowledge with you. Rosy though the transition has seemed, Apple is actually deeply unsatisfied with Intel (I think Steve's dissatisfied with Intel's wretched product names). Because of this, Apple's switching to SPARC. I know, I know, it sounds crazy. But you'd have said the same thing about the Intel switch just last year! SPARC brings us back to the glory of RISC, and it's open source (open source hardware!), that's very hip right now. Best of all, instead of Sun buying Apple a la 1996, Apple will be the savior to the few dozen remaining Sun fans, helping the company limp along for a few more years after the Microsoft money runs out.

...And Here's A Cold Splash Of Reality

In all seriousness though, we have done some thinking about what will be in 10.5, and perhaps the biggest change could relate to Windows, Boot Camp, Parallels, and true virtualization (where Windows apps "just work" on OS X). "Super Rosetta", as Gus put it in his aforementioned post.

If virtualization is part of Leopard, it's interesting as a user and fairly frightening as a developer. While tippling with Daniel Jalkut earlier this week, the topic of virtualization came up and got me thinking. Virtualization would instantly bring a huge number of "Windows" applications to the Mac platform, but there are two very important things it would not do.

1) Virtualization would not initially bring (m)any new users to the platform.

If suddenly a Macintosh computer looks like a Mac and acts like a Mac but also runs Windows software, that's certainly appealing to lots of people. However, meaningful new user growth on the OS X platform from that would take a lot of time. Meanwhile, the existing Mac user base would suddenly have a lot more software to choose from, while developers still had roughly the same size user base to fight over.

2) Virtualization doesn't instantly bring Mac software to Windows.

That seems obvious, but it's still important, because it means that Mac developers see no up front gains from virtualization. I'd also go so far as to say you'd never see Mac apps virtualized on Windows. Unlike Apple bringing Windows software to the Mac, Microsoft would gain very little from having Mac apps on Windows.

So it really seems that virtualization would primarily benefit Mac users and Windows developers. Mac users suddenly have a lot more software overall and many more choices for any one type of application. Windows developers suddenly gain millions of new potential customers. Admittedly, a fair chunk of these people will sneer at the thought of soiling their Macs with Windows software, but many more will not. After that, one must wonder for how long it would be viable for Adobe to produce Windows and Mac versions of Photoshop. Does both the quantity and quality of Macintosh software eventually drop due to virtualization? It certainly seems possible.

In short, if high-quality virtualization (read: as seamless as Rosetta) occurs, developers all over the Mac will be left with a lot of new competition and the same size user base. Over time, that base would (hopefully) increase, but would everyone survive? Would the new users purchase true Mac software or stick with the software they used back on Windows? I don't know, and I don't think anyone can claim he does. I do know that the whole thing is very interesting and more than a little bit scary.

Posted by Paul | Permalink | View/Post Comments (6)

Comments

Alison Scott
Thu Aug 3 07:02:09 2006

I don't think virtualisation is quite as worrying for Mac developers as you think. Yes, it's great to be able to run Windows apps on a Mac, just the same as it's already useful to be able to run Linux programs on the Mac. But these programs are never properly Mac like; they don't quite fit. So when I'm seeking out software I always aim for something designed from the ground up for Mac.

A really good example of this is Firefox. Great browser, but really doesn't work in grain with the Mac. So I'm using Camino, which has many of the same virtues as Firefox but feels entirely more Maclike (and in particular, implements keychain really effectively).

Michael
Thu Aug 3 10:59:02 2006

Your analysis is very good. And history would offer some support to the idea that users would not purchase true Mac software. OS/2 had great virtualization support, and as a result nobody wanted to create apps for it, when they could develop for windows and run on both platforms.

Daniel Jalkut
Thu Aug 3 12:01:36 2006

Instead of using up all of the comment space here, I posted my lengthy reaction to this on my own pages:

http://www.red-sweater.com/blog/167/compete-with-what

Daniel

Paul
Thu Aug 3 12:16:53 2006

Alison - Hopefully that's the case, but without seeing it, I don't want to assume anything. Certainly Linux apps don't fit on the Mac, but Windows apps are a bit different. They may not fit, especially initially, but the Mac may simply change to make them fit (and get worse in the process). Or maybe it would simply be something where they work, but look like crap, and people use it to scrape by when they need a Windows-only application.

Michael - Thank you. It's certainly possible, anyway. I hope that developers on the Mac would have more loyalty than that, but the big players may well not. That was a concern with -Boot Camp-, so it should be an even bigger issue with virtualization.

Daniel and I have been chatting about this, and he raises some great points. He refers to noise being increased, from simply having 50 different apps in a genre (2D vector art apps, for example). I'd consider that part of "competition", but he's been more specific with it. I think (and hope) that he's right, that Mac devs would have a huge advantage, one that wouldn't disappear. But my last two questions loom large - Would everyone survive? Would the new users purchase true Mac software or stick with the software they used back on Windows?

As well, just posted, virtualization for games on the Mac. I guess some of this may play out in a test bed for the rest of us to watch.

http://insidemacgames.com/news/story.php?ID=13850

Shawn Levasseur
Thu Aug 3 14:14:47 2006

I doubt virtuallization will occur on the mac without having to buy a copy of Windows. And that extra purchase (and the security and technological headaches that come with Windows) will be the difference that will give Mac applications a clear edge.

Also, programs within a virtualized system are probably slower than native programs. Nowhere as bad as system emulation (Virtual PC), but slower none the less.

Scott Ellsworth
Thu Aug 3 17:19:03 2006

Assume that PC products work flawlessly on the Mac. Would everyone survive?

No.

Would they have survived against just Mac competition?

No, likely not. It is hard to run a business, and products do die.

I suspect, based on my own buying habits, that Mac developers will do a better job of creating good UIs and feature sets that attract users, and having the PC products on the Mac might just be useful. It will show the Mac developers just what the users want.

For me, the apps that I need to run under windows are: some sequence search software that one of my clients wrote in C#, the Campaign Cartographer gaming mapping tool, the GURPS character generator, MathCAD, and QuickBooks. I might also use it to run Elder Scrolls, or perhaps Starfleet Command.

QuickBooks does run on the Mac, but the file formats are not exactly the same between Mac and Windows, and QB Mac does not offer the same payroll option as QB for Windows. Solve these, and we would just buy a QB/Mac license for her work.

Frankly, Campaign Cartographer kinda bites. It is a complicated as any other piece of CAD software, and thus has a learning curve like a wall. It has very little style, and very little fun to using it, but it sure does produce nice output. Were a Mac company to write decent competition that could do what it did, but with some grace and elan, then CC would get slaughtered, but I doubt a Mac company is entering the market.

The GURPS character generator does have some very good competition. That said, the two tools do different things.

The sequence software may get rewritten, or it may not. That said, once I start work for Google, this becomes less of a concern to me. Such niche products do not drive Mac developers out of the market - they allow users in that niche to use a Mac at all. A good mac package will keep its following, I suspect.

I like MathCAD. It was on the platform once, and will thus not likely come back. Would be nice, of course.

Elder Scrolls sounds like a neat game. The question in my mind is whether I buy it now for my mac, or buy it for PS/3. Were it to run on my mac under virtualization, then I might buy the PC version for my Mac, but even though I had Boot Camp, I waited for Civ 4 Mac, and am waiting for Heroes 5. This says to me that Mac-specific products win, for me.

I want to run Starfleet Command, but that is getting hard to find even for the PC - it is seven years old, after all. I do rather wish it had come out for the Mac.

Heck, if they bring back the old Atomic wargames, I might be busy for months playing them.

So, in summary, it looks like I would find virtualization useful for a few products here and there, but when a competing product does exist for the Mac, the PC products do not win.

Scott

This post is archived, and commenting has been closed.