Daring Fireball: Antennagate Bottom Line
daringfireball.net
daringfireball.net
Whether true or not, the location of the band is particularly ill conceived, as it is touched by holding it in nearly any comfortable position, even with the ring finger when held in the right hand and in the left as demonstrated by Jobs himself.
The fact that the antenna is externally conductive is the real issue here, as it has been demonstrated that antenna performance is significantly improved by placing a piece of duct tape or other insulating material over the band gap. It is very clearly, then, a poor design to not have insulated the antenna in a non-conductive coating.
This is not a logic test that anyone should fail. Does the antenna have a conductivity issue when held in the hand? If yes, can the conductivity issue be resolved with minimal impact on design thereby improving performance? If yes, did Apple implement such a design? If no, isn't Apple guilty of releasing a product with an inherent design flaw?
I think the answer is a resounding yes. I also think other smart phones do not factor into the discussion in a meaningful way.
Yes. Many times.
Just like almost every other engineer, including one hundred percent of those who practice agile methods and release minimum viable products.
Also: Jimmy Carter was guilty of adultery in his heart.
I'm struck by the fact that Apple criticism has acquired this theological tone. It's like Catholics debating the nature of original sin. Were the engineers born in perdition, or did they fall from grace? Can they be redeemed by faith alone, or must they perform good works, and how many additional Hail Marys are necessary?
The bottom line is that this particular issue, whatever its Platonic form, is apparently not serious enough to get me off the iPhone waiting list. I am getting bored with waiting. Return your phones, people!
The other bottom line is that mobile phone discussions used to occasionally stray to other companies besides Apple, other products besides iPhone and other people besides Steve Jobs. Now, apparently, there is nothing better to do but debate the color of Steve's invisible halo. Is this a signifier of the state of the industry right now? At some point, Motorola will release an exploding phone just to get a share of all the press.
So, you don't actually own an iPhone 4, and haven't suffered with this issue, then?
I hope your experience will be much better than mine. Truly.
Notice I never argued that Apple's error was severe enough that the only rational recourse was to return the phone itself. So, perhaps I did; perhaps I didn't. It's just not relevant.
I did not, in fact, return my phone. The bumper has fixed most of the issue, though it is still not an elegant solution. I very much suspect that Apple will have a hardware based fix, in the near future. Even if they don't, I get relatively good signal where I am, and so am not usually affected.
Still, I have had "searching for signal" issues that would go away when I held the phone differently.
-Cheers
Only shareholders, not users.
Except this is entirely the wrong question to ask. More appropriate from an engineering perspective would be: is this an acceptable trade-off? Is there an attenuation issue? Yes. Does it appear, from independent testing, that the iPhone 4's antenna is nonetheless much more sensitive than previous models and much better at holding onto calls with low signal? Yes. These are the two sides of the trade-off, and the only useful question we can ask is whether this trade-off is acceptable.
Because, ultimately, engineering is all about trade-offs. You give up this, you get that, and you aim to make the best trades you can.
"I don’t know if this signal-degradation-in-left-hand thing is really a widespread problem or not yet, but it’s not reasonable to tell people not to hold the phone this way."
- John Gruber (Daring Fireball)
(I was going to guess that it's GSM vs. CDMA issue, but of course, AT&T's 3G network is CDMA.)
The conversation shifted because the speculation was proven to be correct, as stated in the opening paragraph of the piece.
Is this a significant problem? No.
Do all cars have this problem? You bet. Not this exact problem, but still.
Is this a major screw up on part of a company that prides itself on producing high-end polished cars? Totally.
Do their claims that it's not a big deal affect their image? Absolutely.
And regarding the support number of 0.55 percent, not everyone calls support for simple things like a dropped call or no reception, add to that the news coverage where even local news stations are reporting the antenna issues and apple providing cases for free less people are probably willing to call for a known issue.
Now maybe this is just typical Apple PR nonsense to make something as humdrum as product testing sound sexy but if not then it's mind boggling.
New employee: "So where do we test the iPhones radios then?", "We don't", "Oh, okay, so what does the antenna engineer I met at breakfast do in that building over there all day", "Nothing".
Having your system running on intel chips, I can see why that would be kept secret. Keeping your new models secret before launch, fine. Secret anntenna testing for what must be the most famous phone in the world, I don't get. Where is the benefit?
When you are not in contact with the antenna bridge, signal quality and throughput is much higher than any other iPhone. By a large margin.
When you come in contact with the antenna bridge, signal quality degrades considerably.
I'm sure this was discussed at length by many engineers, on many occasions. They did a risk-benefit analysis, and concluded that the pros would outweigh the cons. They probably also weighed the option of trying to engineer something that allows them to have their cake and eat it too, but the need to have something to demo at a keynote, and phones ready for manufacturing didn't make it possible this iteration.
-----
I go through this all the time. "Well, we can take this naive approach. It works. But processing time is O(log n). I think we could take a slick approach and reduce it to a worst case of 2n, but those worst cases are edge cases. Even without this approach our processing time is almost always n anyways. It'll add 10-20% to the development timeline to handle these edge cases gracefully."
The answer always ends up being not to waste your time designing for edge cases that happen in specific (and in our case, controlled and planned) circumstances.