I really like reading these types of articles, so I thought I’d write one myself. So here it is: what I expect will, and what I would like to, be included in the feature set of the next version of Mac OS X. Well, Apple has already announced it will be including a screen reading utility. I'll just quote them:
[Spoken Interface] reads aloud the contents of documents like Web pages, Mail messages, and word processing files; provides a comprehensive audible description of your workspace and all the activities taking place on your computer; and includes a rich set of keyboard commands that allow you to navigate the Mac OS X interface and interact with application and system controls.
What does this mean for us? Well, I doubt these type of facilities will be much use to the majority of Mac OS X users (besides temporary wow value). However, I do expect that the speech voices will be improved—hopefully significantly. (I wouldn't mind if Apple ditched all of those tacky novelty voices, especially.) I know I enjoy using speech to provide feedback in Applescripts. An overhaul of the speech recognition program might get more people using it. More significantly, I believe this will make Applescript GUI scripting a lot more accessible.
Second off the bat is Xgrid. This is a program that builds a cluster of network Macs that share CPU power. Of course, it uses Rendezvous to make the whole process trivial to set up. By the time 10.4 is released I expect Xgrid will be out of beta and included in the box. This may or may not be one of those “keep this quiet because only developers care” features a la X11; there’s a chance they'll be able to work it so well that home networkers will want it also.
Back in my very first column here at billpalmer.net, I wrote something about Windows XP including all sorts of features (three years ago) that Mac OS X is still catching up with. Microsoft’s Remote Desktop Connection is a prime example. This is a program that lets you log into your computer remotely over a network and control it as if it were in front of you, and it is bundled with Windows XP Professional. Apple sells a similar program separately for US$500! Microsoft’s program is real top notch, as well. It’s high time for Apple to include this with Mac OS X.
Well, this one’s kind of obvious. Now, I'm sure that you've all read about Apple’s new program Motion. What was particularly significant about this app, besides its innovative approach to motion graphics, is its incredible performance. Spectators at NAB simply couldn't believe how powerful this program was. How does it do it? Lots and lots of G5 optimisation. Once the programmers at Apple finish up on Motion, I fully expect them to work their same magic where they can inside Mac OS X itself. And I expect them not to focus at all on the G4. If I were Steve, I’d want as big a jump in performance as possible when moving from the aging G4 platform to the new G5 one, to accentuate the improvement as much as possible. (Don’t get me wrong, 10.4 will be more optimised in general. It'll just be more optimised for the G5.)
If you've ever gone past Dave Hyatt’s weblog, you'll know that the development of Safari is anything but stagnant. While public releases are relatively few and far between, the progress the Safari team makes is pretty phenomenal. Recent developments include huge Javascript improvement (better performance than any other browser), not to mention squashing every rendering bug they can find while implementing support for ever more CSS features. By the time 10.4 debuts, I expect big announcements along the lines of “Safari is the fastest, most standards-compliant browser—100% CSS1&2. And look at all these great features…” I wouldn't presume to second guess their announcement, but various standards such as MathML and SVG would be great candidates for inclusion. I would expect as much of CSS3 as possible, too (it’s not yet a finalised standard).
I'm not really talking about new features at all, am I? Not being a User Interface researcher, I'm pretty much all out of ideas with regard to fancy things like Expose and the like. I expect new and amazing things to make the crowd go “ooo” at WWDC, but I'm not the person that can pre-empt them. Suffice to say that I expect more fluid movements and fancy motions, and more improvements to the way things look (the changes in the appearance between Jaguar and Panther were so good it’s a shock to the system to go back and see Jaguar screenshots—imagine what 10.0 would look like to us now!). I wouldn't mind seeing some more faux-3D stuff in the way windows are structured on screen, with subtle perspective effects applied to background windows. It’d be also really nice to have dynamic desktops. These would work in the same way that you can assign screen savers to replace the desktop picture, but be fairly unobtrusive and use few CPU cycles.
Moving on, I want more transparency with how X11 apps behave in Aqua. Separate Dock icons without the X11 environment wrapper and double-click support to open documents in X11 apps would be wonderful. They managed it with OS9, I'm just not sure if they want to do it with X11. X11 apps are fairly alien, after all. That’s a poor argument, though, because you only can’t get an X11 app without knowing what you’re getting yourself in for. Great programs like FontForge and Xemacs would be so much more convenient if they had their own Dock icons.
At the moment, Mac OS X has no provision for scaling the output resolution of its graphics based on physical monitor size. The iBook is a good example here: we have a 12" and a 14" LCD, and the only difference is the size at which they display things; their pixels correspond exactly one-to-one. However, there are many Wintel laptops that have resolutions much much greater than an iBook. What would happen if we tried to use Panther on one of those machines? Our GUI elements would be tiny. Literally miniscule blobs. Luckily, Apple was thinking ahead when they designed Mac OS X and the underlying graphics component, Quartz, is in itself device independent. With some small tweaks it is theoretically possible to scale the raster output from the Quartz compositor to match a larger resolution such that our GUI doesn't shrink to Lilliputian proportions, while at the same time providing greater resolution (and hence detail) for on-screen vector elements, specifically text.
There is an in-depth commentary about this issue in the comments to an article in ondot.net. Suffice to say that our screen controls look good with our 12"–15" 1024x786 LCDs. On a higher resolution screen you can scale the bitmap up in order to keep the same physical size, and even then it is trivial for Apple to provide higher resolution bitmaps for higher resolution monitors. Well, if it’s trivial, it’s overdue! This would allow very higher resolution monitors (think 1920x1200 pixel at 15") and would be a very good selling point against Longhorn, which will offer similar (perhaps improved) functionality. “Actual size” in Preview.app would actually mean something, too.
Towards Mac OS X 10.5
I have purposely left some things unsaid until now because I think that Apple might be working on them—but they won’t be done for a while yet. I’d like to first go back and revisit Xgrid. Consider a home network. This is an app that has the potential of pretty much decentralising computer power in the network. Buy a new computer and rendering times increase exponentially. But consider another NAB announcement: Xsan.
Perhaps it will be possible in the not-so-distant future to have a home network connected with Xgrid and Xsan so that no matter which computer you use, you have access to your home folder as if it were stored on the computer which you’re logged in on. In this scenario, you could have a home server, called an iServe, appropriately enough, which provides the bulk of the storage. Digital recorders could record TV for playback anywhere in the house (the TVs themselves would of course be networked); rendezvous wireless stereo receivers could play your music in HiFi no matter which room you’re in. The iServe and the iSan work together to redundantly backup everything without a hitch. I think it’s a powerful vision that fits in perfectly with the digital hub idea. Hopefully Apple thinks so too!
For my penultimate paragraph, I’d suppose I should talk about metadata. If you don’t know why people expect big things, John Siracusa’s lengthy (as always) metadata article can tell you what metadata is exactly, and you can read some of Scot Hacker’s Tales of a BeOS Refugee to learn what the big deal is all about. There are many accounts about how amazing extensible metadata support in BeOS was. Well, the guy who wrote the BeOS file system is now working at Apple, so there have been many rumours of similar fantastic-ness appearing in Mac OS X. If I had a crystal ball, I’d love to confirm this for you. As it stands, I think there’s probably something planned along those lines—but I have a feeling it’s not for this release. I expect to see the beginnings of something, though. We'll only know later this year. Who know’s what it'll be?
In the end of it all though, I think we know a lot of what we expect. Less bugs, more speed. A couple of fancy visual effects. More efficient apps, and a couple of new ones. Every single year. Isn't it great?
Will Robertson
PS: Here are two other 10.4 wishlists that I've read. Turns out I've got some ideas in common with them. One at codepoetry and the other, older one, at osnews (which may have seeped into my unconscious before I wrote my list).
Switch to the Mac now:
