Of course it came with OS X installed, so the first thing I did was update it, tweak it to my liking and install tablet and printer drivers. That was all painless, and I had a fully functional Mac/X11/BSD system in no time. Buoyed by my success, I decided to install Windows XP. Boot Camp Assistant took a while to do the live repartitioning, but it worked, and I had a FAT32 partition to play with. The machine booted from the XP install CD, and I proceeded to format the partition as NTFS and install Windows. It took a long time, and I had to restart about ten times during the install, update and driver install process, but in the end I had a working Windows installation, too, and I could switch between OS X and XP on boot. That was enough for one night.

I downloaded and burned the Fedora 10 x64 live CD, as it seemed like the path of least resistance. That was where my luck ran out. Trying to create a partition for Linux wasn’t anywhere near as smooth as creating one for Windows. Using Disk Utility while booted from the hard disk went nowhere. Apparently the live partitioning feature isn’t that stable. Booting from the OS X install DVD allowed me to repartition the disk, though.

Next, I tried restarting from the Fedora live CD, but no joy. Didn’t get past the firmware’s OS selection menu. Switching the machine off and booting from the CD worked, though. But then I was left with just a blinking cursor — X11 wouldn’t start. I realise that they won’t put the NVIDIA drivers on the CD because they aren’t “free” (as in beards), but shouldn’t it at least be able to choose a VESA driver for compatibility in the event that it doesn’t recognise your hardware? Windows manages to do that. Lack of X11 didn’t stop me switching to a console to do stuff, though. So I logged in and used parted to reformat my Linux partition as ext3, and then created a contiguous 2 GB file to use for swap space.

Since I’d played with the partition map, I had to fix BOOT.INI for Windows. But now I had a problem: despite correctly identifying the GPT partition scheme, and not altering the layout at all, parted had somehow managed to damage the GPT in such a way that only one partition would show up in any OS X utilities, and attempting to verify the OS X boot partition with Disk Utility caused the system to hang. It also screwed up the fake MBR so Windows wouldn’t boot at all. After several hours of messing around with partition table editors, I did manage to fix it without losing anything, though. Then I booted to a Windows recovery console (using the install CD) and fixed up BOOT.INI (by incrementing the partition number). I was back to having working OS X and Windows, with the dubious benefit of also having a blank ext3 partition. By this stage, I was pretty pissed off with Linux and considering installing Solaris just to make myself feel better. But it was time to sleep again.

Thinking about this the next morning, I realised that I now knew how to fix the partition map if I managed to screw it again. This emboldened me enough to give Linux another shot. (Yet another example of why I should I shouldn’t think so much.) I booted from the live CD and got the console up again. This time I hacked up the X11 configuration to force the VESA driver. That got me a desktop. I noticed that my tablet was working properly (take that, stupid Ubuntu live CD), and started the installer. It had to format my ext3 partition again (so much for the contiguous swap file), but it seemed to go through its thing. Then at the very end, it told me that the install had failed. It hadn’t screwed anything up this time, though. So I had a system with a working OS X installation, a working Windows installation, a working Windows installation, and a non-working Linux installation.

On the advice of other Fedora users, I downloaded and burned the Fedora net install CD. Booting from this one got me X11 using a VESA driver straight away. However, my tablet didn’t work at all. What’s with that? It provides an HID pointing device interface for compatibility that OS X, Windows and even the boot menu in the Mac firmware can use, but Linux can’t. But anyway, I did manage to work my way through the installation process using the keyboard. It took quite a while to download everything, but it didn’t fail at the end this time, and it did yield a working installation.

After booting from my new Linux installation, I was pleased to see that it was using a working VESA driver for X11 video, and the tablet was working. It also recognised my external RAID, and mounted HFS+ and NTFS partitions. It was quite easy to configure RPM Fusion repositories and install the proper NVIDIA drivers. That got me higher resolution and better performance. I had to install the glibc compatibility package before I could install my printer’s drivers, but that was easy, too. So in the end, I got the lovely triple-boot system that I’d imagined.

But despite the fact that it worked quite well in the end, it wasn’t a simple process, and definitely wasn’t something a typical user could have achieved. Considering how unreliable the live CD is, it should be marked as unstable or experimental, and users should be steered away from it. If parted can’t deal with GPT, it should give you some kind of warning before it trashes your partition map. A common graphics tablet should at least give you basic pointer functionality. Bloat like GJC (which is completely broken), the Evolution mail client and Gutenprint/foomatic (only needed if you have a cheapo printer with no PostScript) shouldn’t be part of a default install.

Fifteen years ago, you needed to be technically adept to install any operating system. If you weren’t good with computers, you’d call your technically inclined friend or relative to help you out. In a way, no operating system was really “ready for the desktop” back then. Linux feels like it’s still at that point. Once it’s up and running, it’s fine, but to get it there, you need considerable problem-solving skills and/or friends to help you out. Windows and OS X have got past that point — anyone could install either of them. Linux really does have some serious catching up to do.

I’m with Hutchison/3 in Australia, and they have provide three packet data services: MMS service (pay per message sent – unmetered data), the “walled garden” (pay based on services used – unmetered data) and the wild Internet (pay based on usage). I use the “walled garden” quite a bit – I get free news an weather, so I read it on my way to and from work. I also pay a flat fee for unlimited e-mail use. I occasionally use the expensive Internet service when I’m somewhere out and about and I decide I need to get online for something. When I do, the phone looks like a Bluetooth modem to the computer, and I can use the connection as I wish (provided I pay the bill at the end of the month). I can also use the Internet service with applications that run on the phone, but I don’t do that very often.

Now if I had unlimited data usage included in my contract, I’d probably feel a need to be online while out and about a lot more. Let’s face it: when something’s free, we use it more. But if everyone did this, the networks would be brought to their knees by the huge amount of traffic. So AT&T need some way to stop iPhone owners from actually taking advantage of the “unlimited data” in their contracts.

The way they’ve done this appears to be by making a deal with Apple to cripple the iPhone: it won’t work as a Bluetooth or USB modem, and Apple won’t allow applications that would generate lots of traffic on the cellular network. So VoIP applications are only allowed on WiFi, and they’ve pulled Nullriver NetShare, an application that allows the iPhone to act as a modem, from the application store (to be fair, Nullriver was asking for trouble, considering their previous claim to fame was an installer for putting bootleg applications on “jailbroken” iPhones).

But the rest of the world doesn’t have unlimited data usage. We pay for the data we send and receive over the air. So why can’t we use it as we wish? It’s a bit unfair to restrict everyone to make AT&T happy.

In Leopard, Input Managers are no longer supported, and are severely restricted. Now I know why this is being done – there is great potential for Input Manager malware. Also, Input Managers were never suitable for system-wide input because they didn’t work with Carbon applications. But I’m sad to see them go. Mac-style Input Method components are a far less elegant way of performing the same task (albeit with far lower potential for evil), and the APIs Apple themselves use for writing new-style input methods still don’t seem to be documented on ADC.

The other thing that’s disappeared is NetInfo. It’s been replaced by Directory Services. I guess it’s time for us to learn to configure static hostname resolution, DHCP/BootP/NetBoot servers, unusual account settings, and all the rest of it all over again.

• I like having a mechanical keypad – say what you like, but you can’t use a virtual keypad without looking at it.
• I like being able to use Java MIDP applications.
• I can’t tolerate the lack of performance you get without 3G.

The iPhone isn’t a smartphone – it’s a feature phone. The defining feature of a smartphone is the ability to run user-installed applications, and the iPhone forbids that (it’s priced like a smartphone, though). Now if you want a smartphone, buy one – there’s no point buying an iPhone and trying to hack it. Apple doesn’t care about people who hack iPhones – they already have your money.

I have no issues with the price drop – that’s standard practice for mobile handsets. You make as much money as you can to begin with, and then lower the price to keep sales up. Everyone does it. If you complained about that, you’re a crybaby. What I really don’t like is the amount of crap coming from Jobs and the fanboys.

First of all, the argument that you’re better off without user-installed applications because they could bring down a mobile network. Well all I can say is that it would have to be a very flaky network for that to happen. We’ve had mobile applications for Java MIDP, Symbian, Windows Mobile and BREW for years and there hasn’t been any trouble. Mobile networks are designed so that a badly behaved device won’t bring them down, and mobile handsets are designed so that one bad application won’t bring them down. And if that was the real reason, the iPod Touch would allow you to install applications. No, the real reason is presumably due to the way Apple have got a deal where they get a cut of revenue from the mobile carriers – if you could install your own applications, it could erode this.

Next up, all the crap about why 3G isn’t any better. It’s true that early 3G handsets had poor battery life, but they’re much better now, and at least as good as EDGE handsets. But even if 3G really did give poor battery life, it would be worth it. The data rates and round trip times on 3G leave EDGE for dead. EDGE is a hack on top of a hack – GSM wasn’t designed to provide packet-switched data. Also, voice call quality is far better on 3G. It sounds clearer, and sounds great up until the point where it drops out. Contrast this with GSM where the conversation starts to break up and get noisy long before you lose the connection. And last but not least, you can’t roam in Japan without 3G.

The very aggressive SIM locking is a bit of an interesting issue. Now the reason for it is obvious – Apple gets a cut of the revenue from the carrier, so they want you to stay. None of the other handset manufacturers have deals like this, so they don’t care anywhere near as much about whether you can unlock your handset. You may be able to get away with this practice in the US, but EU regulations clearly state that carriers must allow customers to buy their way out of contracts and unlock their handsets if they do. If they’re perfectly fair and balanced, they will enforce this, and compel Apple to provide a means of unlocking iPhones. Let’s hope they don’t have to be dragged through the courts first.

Last of all, there are ringtones. I do have a variety of ringtones on my handset. I assign different tones to different contacts, so I can tell who’s calling immediately (although I have it set to vibrate only most of the time, anyway). I didn’t pay for any of my ringtones. My handset will accept any AAC, MP3 or MIDI file as a ringtone. Assuming I am not violating copyright by possessing the files to begin with, I am entitled to use them as ringtones. I have no idea how anyone can say that iPhone ringtones are a good deal: paying for a song twice or more so you can hear it when someone calls is just stupid.

• I think the new keyboard is just plain ugly, but that’s purely subjective. I think the PowerBook keyboard feels better than the MacBook keyboard, too, and would have preferred a desktop keyboard based on that.
• The wireless version is obviously designed for using on your lap, rather than on a desk. The lack of a numeric keypad allows you to have it both physically and logically centred on your lap.
• Dashboard functions/exposé on the left are interesting – here’s my theory: on a PowerBook, it makes sense to have it on the right, since you need your left hand to get to the fn key while you hit the F9/F10/F11 key with your right hand (otherwise you get keyboard illumination control); however, on the desktop keyboard, you need to use your right hand to get the fn key, so it makes sense to have the multiplexed F-keys on the left-hand side of the keyboard.
• I was sad to see the help key fall into disuse, and I’m sad to see it ultimately disappear.
• I won’t get one of these keyboards. I don’t like wireless input devices that need batteries (hence by Wacom Intuos3 with wireless power to the pen and hamster – it can’t be a mouse without a tail). I’m also very happy with my Sanwa IceKey keyboard, which has very nice notebook-like key mechanisms.

OK, that’s enough random garbage for today…

A properly maintained iPhone battery is designed to retain up to 80% of its original capacity after 400 full charge and discharge cycles. You may choose to replace your battery when it no longer holds sufficient charge to meet your needs.

Now Sullivan has decided that this means the battery becomes completely unusable after 400 cycles, and written a whole article about it. So yes, he’s guilty of spreading FUD, and I guess you can call him a jackass for that.

But Gruber has gone to the opposite extreme. Notice the use of “properly maintained,” “designed to” and “up to” in that quote. Gruber seems to think that this means the iPhone battery will hold 80% of its original capacity after 400 cycles. It doesn’t say that at all. It also doesn’t say how much of the 80% capacity that’s retained might actually be usable. Remember, too, that because we carry our phones around with us, we just charge them when we can. None of us are really in a position to “properly maintain” their batteries. In this case, it would appear that Gruber is spreading fanboyism.

Is the built-in battery really such a big deal? To me it is. I like to be able to carry a fully charged spare battery and swap it in. I wouldn’t be able to do that with an iPhone. But it’s a well-known fact that iPhone batteries are non-removable, and the performance figures and price don’t really look that much worse than the competition. So for people all the people who only use one phone battery anyway (I imagine this is the majority case), it probably isn’t an issue.

Suppose I’m a bot herder (I hope it’s obvious that this is purely hypothetical). I unleash malware that takes over computers and “calls home,” allowing me to send spam for my paying clients. Since the number of people who read spam is very low, and the number of people who buy products advertised in spam is even lower, I need to send huge volumes of spam to make my services worthwhile for my clients. And sending lots of spam requires lots of compromised computers.

As software vendors patch vulnerabilities in their software, I have to find new vulnerabilities and write new malware to exploit them. This requires considerable effort on my part, and takes away from time I could spend doing things I enjoy. Also, as more users become more security-conscious, there are less machines left open to attack.

Suppose for a moment Windows, Linux and Mac OS X are all equally exploitable, and writing a piece of malware for each takes the same amount of time. What am I going to do? Am I going to write three sets of malware to attack the three platforms, or will I pick one to concentrate on?

Of course, the answer depends on market share. The more even the market share, the more likely I would be to write malware for multiple platforms. Also, it’s worth thinking about where the machines are primarily used.

Linux is used primarily in server and professional environments. Machines that are critical for business operations run by tech-savy operators means the machines are more likely to be secured properly and suspicious software will be removed promptly. So scratch Linux. I want to target home users with DSL or cable internet.

So I’m left with a choice of targeting Windows or OS X. What do I do? I look at market share. I know these figures are probably wrong, but suppose OS X runs on 5% of my target machines and Windows runs on 90%. What am I going to target?

The answer should be obvious. I’ll target Windows. I could target OS X as well, but then I’d be spending twice as much time writing malware for less than 6% more compromised machines to send spam from. It just doesn’t make business sense.

First of all, the bad:

• Intel GMA950 graphics chip – this gives me all kinds of trouble with external monitors. After using iMovie, it won’t detect external monitors until I restart. After waking from display sleep, external DVI monitors can lose synchronisation with the video signal and need to be disconnected and reconnected (this messes up window positions). The performance isn’t bad, though.
• Reflections off the glossy screen – when that happens, it ranges from really distracting to just plain unusable. Matte screens are never that bad. Sure, it looks great when the light is OK, but it’s all or nothing.
• No internal modem – hey, people still like to get faxes, so I want to be able to send them. Sure, the USB modem is funky, but it’s one more thing to buy and carry around.
• Only FireWire 400 (IEEE1394a) – if you’re only giving us one FireWire port, can’t you make it a fast one? We can always use adaptors for slower devices.
• MDVI video output – this wouldn’t be a problem if they included adaptors, but they expect me to go and spend another $100 or so to be able to plug in to external monitors and TVs.
• No 802.11a – like it or not, Steve, 802.11a is far more efficient than 802.11g. I know the number is the same (54 Mb/s for both), but there’s a lot more to it than that.
• Temperature – this thing gets a lot hotter than my wife’s old iBook G3.
• Mac OS X 10.4 “Tiger” – Spotlight is useless, Dashboard is a waste of memory, the font rendering is nowhere near as good as Panther and the input menu behaviour is really weird. How does it order the items in the menu, and why does Kotoeri keep deciding to change the underlying keyboard layout for kana input mode? I liked old, mbox-based Mail.app with the mailbox drawer, too.
• Apple Remote – it’s a useless piece of junk. Why couldn’t they include some video adaptors or a modem instead?

And to balance that, all the really good things:

• The case – I like not having a latch (like the original iBook), the keyboard is far better than the iBook keyboard, it’s small and light and it looks great.
• MagSafe power connector – why didn’t anyone think of this before? The is one of the best things about the new Apple notebooks.
• The glossy screen – white on black text in terminal windows is a thing of beauty and arcade games in SDLMAME look so good! The brightness and contrast are excellent. Too bad about the reflections, though.
• Dual CPU cores – it’s great to have the system loaded down and still have a really responsive UI. Two physical CPU cores really helps with that.
• Battery life – it’s very good if you aren’t doing anything too heavy. But they’ve jammed a pretty big battery in to get there.
• Gigabit Ethernet – it’s about time that was standard on low-end machines.
• Video spanning – fixing another silly limitation on the iBooks. Spanning is far more useful than mirroring most of the time.
• The video camera – at first I couldn’t get excited about this, after all I can make video calls on my mobile phone, but 3G coverage isn’t available everywhere, so it’s occasionally useful.
• Mac OS X 10.4 “Tiger” – yes, it makes both lists. Xcode just keeps getting better, Rosetta is brilliant, Tiger is faster than Panther overall, you don’t notice that this is the first release on completely different hardware, Jabber in iChat let me get rid of another IM program and there are some very nice additions like the ability to re-map modifier keys without hacks.

Now before I say anything else, I just have to say that if you develop for OS X and you haven’t used Shark, you don’t know what you’re missing. It’s performance profiling how it should be done: simple to use and easy to understand the output. Any non-trivial application could do with optimisation. Even if CPUs are fast enough to run slow code at acceptable speeds, users appreciate snappy applications. And what’s to say they don’t want to use spare cycles for something else? I’m compiling SDLMAME in the background while I blog, for example. And CPUs use less power when idle, so notebook batteries last longer with more optimised applications.

Anyway, getting back to the topic at hand, using Shark was a very interesting experience. I was quite surprised at some of my findings.

First of all, on the topic of libxml2 (the Gnome XML library). Now libxml2 is really cool. It can read an entire XML file into a friendly structure that you can walk forwards and backwards, and edit and even write out as an XML file again. It also validates the document as it goes. I wasn’t using it for any of these reasons, though. I was using it because it’s included with OS X so I could conveniently dynamically link to it. But libxml2 was turning out to a major performance bottleneck in my application. The interesting thing was it was taking about twice as long to deallocate an XML document structure as it took to pare the 24 MB XML file in the first place!

Now my first thought was something along the lines of, “Gee, garbage collection really is a good idea!” If you think about it, garbage collection would let you completely side-step the deallocation step: when the heap is swapped, there are no references so the objects aren’t copied. Using a custom allocator in libxml2 could solve the problem, too. But I side-stepped the issue in a different way: I switched to Expat which doesn’t build a structure from the document to begin with. The trouble with that is my application has now grown by a few hundred kilobytes because of the statically linked XML parsing code.

The second think is that Core Foundation is considerably faster than Cocoa when it should be doing the same thing. In critical areas, I could get a speed gain of twenty to fifty percent by casting my Cocoa collection objects to Core Foundation objects and calling the C APIs.

So where’s the catch? Core Foundation is less forgiving than Cocoa. For example, if you try to get an value from a nil dictionary object with Cocoa, you just get nil. Core Foundation will crash. Core Foundation also has no concept of autorelease pools (but that won’t stop you from casting a CFTypeRef to id and autoreleasing it).

All in all, I can’t recommend performance profiling enough. It’s always interesting, and often surprising. And admit it, everyone loves trying to make stuff go faster

And integrated graphics chipsets that share system RAM with the CPU are going to suck worse with Mac OS X than they do with Windows. OS X uses the GPU for simple tasks like the final composition of the windows on the screen, and Dashboard’s funky effects, and Exposé, and lots of other things. Every time something like this happens, the GPU and CPU will fight for precious memory bandwidth. This will really slow the MacBook and Mac Mini down.

Please, Apple, can you give us a machine that isn’t top-end, but doesn’t just plain suck? And while I’m at it, can we have 802.11a as well? You know it’s faster and has better range!