Apple’s New Models

It’s a strange time to be raising prices, but Apple has just done it. The new Mac Pros are priced at $2499 vs. $2699 for the older model, but the older model had 8 cores, while the model that “replaces” it has 4 cores. Apple’s website is conspicuously silent on the relative performance of the two obviously comparable model, preferring instead to show that the new top-end model is over twice as fast as the old top-end model. That’s nice, but the new base model is only a shade cheaper than the old mid-range model, and probably (we can’t tell) slower, while the old base model (which was $2199) is gone.

There is some good news. Apple has finally put the nVidia 9400M chipset into the Mac Mini, so the Mini is truly a headless MacBook. This means that the Mini’s performance for games should be, at least, acceptable, and makes the Mini surprisingly attractive.

The one thing that might explain the new Mac Pro pricing is if Apple is opening a space for the long-wished-for xMac. In terms of headless Macs, there’s the $599/799 Mac Mini then there’s nothing until you get to the $2499 Mac Pro. I might add that while the Mac Pro has the new “i7″ architecture chip (i.e. hyperthreading support, which treats each core as two virtual CPUs) none of the new Mac Minis and iMacs do.

In iMac land, things are somewhat, although less, strange. The base iMac used to feature an integrated CPU (I think it was the good old GMA 950), now the bottom two iMacs (20″ and 24” with 2.66GHz CPU) have the 9400M chipset (which is integrated but much better). It’s an interesting change where the $1499 price point has gone forwards in screen size and backwards in GPU quality. Or something.

So, in summary, the Mac Mini now seems like a very attractive machine. The base iMac offers some minor performance tweaks relative to the Mini, but (especially if you already have a monitor) the Mini is very close.

The Times Have Changed

It seems to me that Apple is running the risk of seriously alienating a lot of its customers with its continuing high price points. In a time when most of the computer manufacturers have moved from pushing $1000 configurations to $500 configurations, Apple persists in selling an entry level desktop for $1200. Because Macs have longer useful lives than PCs, many Apple customers who bought (then) reasonably priced Apple systems will soon be upgrading in a world where $1200 buys you a lot of PC but not very much Mac.

It’s worth noting that the pricing on Core i7 (“Nehalem”) CPUs seems to be an attempt by Intel to gouge early adopters which is merely being passed on by Apple (and Dell). I imagine that as CPU prices drop, Apple’s Mac Pro prices will fall (or its base model will gain a CPU) accordingly.

Tabbed Titlebars

Technically speaking, Safari 4 is probably the best browser you can get. No big surprise, of course, since in my opinion Safari 3.x (especially if you use Webkit Nightly Builds) was already the best browser I could find. But Safari 4 has an interesting UI decision in it: tabbed window titlebars.

Safari’s implementation is conceptually very similar to Google’s Chrome browser, but Apple being Apple it’s far more polished, although not perfect.

Logically speaking, tabbed browsers make no sense, and of the different approaches taken, Safari 3.x’s made the least sense (the tabs appeared to be selecting the area above the tab panel, not below).

The fundamental problem is simple: the tab metaphor is supposed to represent selecting from among sections of a set of files or pages using the tabs as reference points. Selecting a tab brings that section to the front and puts the others behind (which isn’t how the real world works, but more convenient). Based on the metaphor, anything not “contained” in the tab is “global” while anything not “contained” in the tab is part of the set of stuff contained “in that tab”.

This basic rule is often violated, but usually its in the forgivable direction of including global stuff in the tab area — this usually simplifies layouts. In browsers, however, the rules are hopelessly bizarre. E.g. the address bar in Firefox is above (outside) the tab selector, but changes when you change tabs.

As I said, Safari 3.x has the silliest tab interface. The tabs are above the web page you’re looking at but visually appear to contain the stuff above the tabs (i.e. bookmarks and the address and search fields), so when you click a tab to select the page it’s related to, it appears that you’re selecting a different set of bookmarks, not a different web page.

Safari 4 has made the window titlebar into a tab selector. One obvious reason not to do this is exemplified by Google’s Chrome — it could look really ugly. But aesthetically Apple has pulled off this design about as well as you could possibly hope. Because the title bar is ordinarily used to move windows around, this is now the default behavior of tabs when clicked and dragged. To move a tab around you use the resize affordance in its top-right — this is a horrible kludge, violating the “if it looks the same it should work the same; it it works different it should look different”. In my opinion Apple needs a new affordance to serve this row, perhaps a two-way or four-way arrow. The tabs now correctly sit above the address field and web page content, but also above the search field and bookmarks.

The new design has some obvious wins: a Safari window now wastes less vertical real estate, and the containment hierarchy is logical.

I believe I’ve said this before, but in my opinion the way tabs work in Google’s Chrome (and now in Safari 4) — modulo fixing the affordance for moving tabs — should be implemented at OS level as a way of handling multiple, docked windows gracefully. Ideally, I could link a Pages document together with Photoshop, say, and tab between them when I needed to find a photo to drag into my document.

Picsel sues Apple over iPhone Graphics Update

Another year, another lawsuit. I assume that Picsel’s lawsuit involves either this patent or this one. Here’s a list of Picsel’s issued patents in case you’re interested.

I’d never heard of Picsel before (their logo looks like something drawn by a bored intern using PowerPoint, so they’re clearly a company lacking in taste) but given their client list (most major consumer electronics companies and most major cellphone manufacturers) they’re obviously not simply patent trolls.

I am not a lawyer, let alone a patent lawyer, and I find the way patents describe processes to be amazingly annoying. The one thing you won’t find in a patent is a vaguely straightforward explanation of what the gizmo actually does. But you can piece it together…

As I read it, the first patent basically involves redrawing the screen using an approximation / cached tiles of the expected display and then replacing it with the actual pixels when they’re available (something pretty much every OS and many applications do to some extent, which seems to me to fall under the “bleeding obvious to anyone in the industry” category — unless Apple for some reason is using the very specific tiling technique described, which seems to me less obvious and less likely), while the latter looks like Picsel attempting to patent something vaguely reminiscent of Quartz Compositor (and hence falls under the “prior art” category since Apple has had Quartz Compositor since long before the patent was issued).

So, I’m guessing there’s at least some merit to the lawsuit and Apple will pay something in the tens or hundreds of millions. (After all, they settled with Burst for $10M and their case was a joke.)

Three New Things, And One More (Of Course)

Just some quick reactions to the Macworld Expo announcements.

iTunes. Well, Apple didn’t announce a new Mac Mini with a 9400M GPU, and it didn’t announce a new bigger or cheaper or somehow more compelling AppleTV. And it didn’t announce any new iPods or iPhones. But the iTunes announcement is probably going to turn out to be more significant than anything else — see the second item:

  • New pricing model. $0.69, $0.99, or $1.29 per song. This is not “pay more for 256 kbps”, but allowing Music labels to charge more for new shiny stuff and less for back catalogue.
  • Everything is going to be DRM free. (8M songs now; 10M soon.) In other words, the recording industry idiots have finally gotten a clue.
  • iPhone can now download music over cellular networks (not just Wi-Fi).

It doesn’t say whether our existing purchases will be stripped of their DRM though. I hope that devil is in the details.

The 17″ Macbook Pro looks great. I won’t buy one. When I get a notebook I’ve learned that its single most compelling feature is being small. (As small as possible without becoming dysfunctional.) There’s a perfect MacBook for me already, and it’s the new MacBook Air (which Apple quietly upgraded to the nVidia 9400M sometime in the last couple of months). Oh, and the new 17″ MacBook Pro has a non-removable battery which, apparently, gives you 8h (if you don’t use the faster GPU), and has a 5y lifespan — 1000 charge cycles — but, as already mentioned, isn’t removable.

iWork ’09 looks very compelling. The key missing features appear to have all been added (except for automatic indexing in Pages and perhaps pivot options in Numbers). It would be nice if Apple released a Tables database component but I guess that would make Filemaker’s Bento look silly. Wait, it already looks silly.

I wonder why Apple didn’t try to integrate Google documents into iWork rather than or instead of doing Do they really expect to become a major profit center? If not, why not simply leverage something very good that does much the same thing that happens to be something Google is doing as a loss leader. could be great, but how much better than Google docs will it have to be to make up for not being free?

iLife ’09 looks equally compelling. I’m one of the people who happens to like iMovie ’08 (I think we’re the silent majority). If you want to create an actual movie, iMovie ’08 is hopeless (mainly owing to poor audio functionality), but then so is iMovie ’06 (for much the same reasons). For cutting together a bunch of footage into something halfway decent in nothing flat, iMovie ’09 looks like it will let us have our cake and eat it. If the face recognition stuff in iPhoto ’09 is halfway decent it will be a huge, huge feature. Music lessons in GarageBand seem like a killer feature, but it really depends on how well it’s done.

iPhone Development

After over six months, I’ve finally got my act together (and waited for Apple to get its act together) and can build my own iPhone apps. So far I’m using Unity’s iPhone Advanced tool rather than the “bare” SDK (Unity essentially builds an iPhone project for you, you still need to build the final app in XCode).

Using Apple’s tools, and I don’t think I’m violating any remaining components of the NDA by saying this, makes it clear that the SDK was released on a highly accelerated schedule. The amount of silliness involved in getting “Hello World” working on your iPhone is pretty amazing. It’s all a consequence of Apple wanting to make the iPhone world as safe and secure as possible.

In essence you have to jump backwards through a bunch of hoops to produce encrypted digital signing certificates and registering all your developers, testers, and so forth and their iPhones and iPod Touches before you can do anything. I won’t go into gory details because it probably is covered by the vestiges of the NDA and it’s boring, but take my word that the process involves a lot of non-obvious (even with a step-by-step checklist) steps that involve telling your left hand what your right hand is doing (i.e. stuff that should be automatic). It’s all very reminiscent of using Lotus Notes. (If you don’t know me, that’s worse than comparing it to Blender 2.3.)

The purpose of all this is to prevent people from installing “any old app” on their iPhone, and to allow Apple to flip a “kill switch” and disable any app which proves to be toxic to users (e.g. steals/destroys data or violates user privacy) or violates Apple’s unstated rules (e.g. don’t build anything vaguely related to web browsing or email) or Apple’s relationship with AT&T (e.g. don’t facilitate people using their iPhone as a cellular modem/base station). I have no objection to the purpose served by all this annoying cruft, but I do think the annoying cruft should be much easier to handle. Basically, all I should need to do is associate my iPhone with my developer account and click OK to have all the necessary crap stored in my keychain and I am done. Instead it’s 25 steps disguised as 12.

Unity’s iPhone tool is simply amazing. (This is amazing above and beyond Unity itself, which is plenty amazing.) During testing you can run your app in the Unity IDE as normal and use your iPhone as a tethered controller/display — Unity sends compressed video to the iPhone and receives the iPhone’s state over USB via a small app that comes with the dev tools. The only downside to testing this way is that your app is running on your Mac, so you don’t see the actual performance you’ll get on the iPhone, and the video can be a little artifacty (woohoo new word!).

So far it looks like you can have around 7,000 triangles visible and still get action game performance, or up to 25,000 triangles for more sedate games. The shader support on the iPhone is limited so blowing out the video hardware’s performance with fancy shaders isn’t really an option, so expect to see a lot of lightmapped scenery.

Anyway, MANTA (which I conceived of as an iPhone game) weighs in at 100-150k triangles visible on screen right now, so it’s not going to be running on the iPhone without some major surgery. I would still like to get it out this year, but my original planned Nov 30 release is impossible. (I’ll try to release the Mac version on time.)