Comments on: Eerie iPhone

By: Sasha

Sasha — Thu, 24 Sep 2009 08:59:57 +0000

A couple things about reception problems with the iPhone:
1) Overall, AT&T was -completely- unprepared for the flood of new users, and even more so the flood of data traffic on their network, that came with iPhone exclusivity. Since the introduction, the company has still not built up their network sufficiently to handle the traffic.

2) One way AT&T has been trying to deal with the service problems is to replace the older EDGE towers with GSM towers. Although this increases their 3g capacity of the network, it has actually hurt network reliability: EDGE towers have greater range and interference robustness than GSM, but transmit data at a slower rate; by yanking out the EDGE towers, they have taken away redundancy. Additionally, 1st-generation iPhones (like yours) are EDGE-only, so over the last couple years AT&T has been slowly removing the pieces of the network that you rely on.

I love my iPhone, but I sure do hate AT&T, and would switch networks in a heartbeat if I could.

By: sirhcton

sirhcton — Wed, 23 Sep 2009 20:13:18 +0000

Don’t be surprised if your message and calling problems are the network, rather than your phone.

By: Marion Delgado

Marion Delgado — Wed, 23 Sep 2009 05:14:51 +0000

It’s a good thing Steve Jobs got his liver and survived so far, is all I’ll say. Then again, Sheril (and her phone) would be Keynote speaker at Macworld 2010 if he hadn’t.

By: Sheril Kirshenbaum

Sheril Kirshenbaum — Wed, 23 Sep 2009 04:45:41 +0000

@13 Marion Delgado,
Totally.

@12 Cain,
I’d only be interested if it might raise money for charity, but doubt I could compete with Marion’s ‘ghost in a jar’ example.

@14 Anthony McCarthy,
Those are still around?

Thanks again to Erasmussimo, Lorenz, and Walker for insight on the trouble.

By: Anthony McCarthy

Anthony McCarthy — Wed, 23 Sep 2009 03:01:14 +0000

Reading down this thread, I’m sticking to my ancient Bell desk phone with a wire in the wall.

By: Marion Delgado

Marion Delgado — Wed, 23 Sep 2009 02:33:05 +0000

I can see the blog headlines now:
“Accommodationism leads to eBay woo”
“Unscientific America author turns to eBay spiritualism”
On the other hand:
“Were pictures from beyond taken with haunted iPhone?”
“Chilling predictions for 2010 made by voice in iPhone”

By: Cain

Cain — Wed, 23 Sep 2009 02:00:46 +0000

SK, If you’re planning on putting a haunted iPhone on ebay let us know, I’ll bid.

By: Erasmussimo

Erasmussimo — Wed, 23 Sep 2009 01:52:40 +0000

Yep, stack crashes are really hard to generate these days, but they’re easier to explain than the messier garbage collection problems.

By: Lorenz Cuno Klopfenstein

Lorenz Cuno Klopfenstein — Tue, 22 Sep 2009 20:35:40 +0000

@Erasmussimo: I agree, it might be an easter egg (perhaps some particular text sequence was entered?).
Memory corruption on the other hand, while possible, is unlikely (and would suggest a bug in the low level memory handling functions of the OS) and the stack-thing is completely impossible: the stack is private to each single process and is – in these days – never programmed explicitly. There is not way another process can mess up another process’ memory and inconsistencies would point to a compiler/runtime bug (also unlikely).

By: Erasmussimo

Erasmussimo — Tue, 22 Sep 2009 19:29:47 +0000

Oh, there are lots of those. The most common are stack errors. While a program is running, it stores lots of temporary variables onto a stack. Let’s listen in on Mr. CPU(NotePad) at work:

Ho, ho, ho. I’ll just push the X-value onto the stack for when I need it later.
The stack is now 1 piece high.
Now I’ve got the Y-value. I’ll push that onto the stack, too.
The stack is now 2 pieces high.
Now I’ve got the Z-value. I’ll push that onto the stack too.
The stack is now 3 pieces high.
Now I’m busy with some other things.
OK, now I’m ready to use X, Y, and Z.
I’ll grab the piece at the top of the stack; that’s Z.
The stack is now 2 pieces high.
Next I’ll grab the piece at the current top of the stack: that’s Y
The stack is now 1 piece high.
Lastly, I’ll grab the piece at the current top of the stack: that’s X.
The stack is now empty.
Now I add X, Y, and Z, and I’m done.

Piece ‘o cake, right? But what happens if Mr. CPU(CrazyGame) jumps in while Mr. CPU(NotePad) is asleep (these things happen in a multitasking processor. Now, Mr. CPU(CrazyGame) is expected to play fair and leave the stack exactly the way he found it, but let’s say he’s a little weird (he was programmed by a 16-year old high school student who doesn’t follow all the rules of good programming) and he gets sloppy and accidentally leaves the stack one piece shorter than it was when he came in. In other words, the stack had three things on it (X, Y, and Z) that Mr. CPU(NotePad) was using, but Mr. CPU(CrazyGame) left only two things on it: X and Y. Now when Mr. CPU)NotePad comes back to use it, he grabs the top item on the stack, which he thinks is Z, but is really Y. Then he grabs the next top item on the stack, which he thinks is Y, but is really X. Then he grabs the next item on the stack — wait a minute, the stack is empty!!! This can’t be happening! This is impossible! Danger, Will Robinson!

Now, the guy who programmed NotePad is a responsible programmer and put in some test code just in case some idiot screws up the stack, which is highly unlikely. But in your case, it actually happened, so now you get his error message.

That’s how it COULD have happened. I’m engaging in total speculation here. There are a million ways this could happen. This is just one possibility. All of which leads to my second rule of programming: When in doubt, shoot the programmer.