Month: February 2013

Jacqui Cheng, of Ars Technica goes mythbusting:

“Apple Makes New Employees Work on Fake Products Until Apple Can Trust Them”, blared a headline—and many others like it—last January. In the Apple-watching world, it has since become common wisdom that the company assigns new engineers to “fake” projects in order to test their loyalty—that is, their propensity to leak—before giving them actual work. […]

But is it true? I was prompted to look into the question after several friends—Apple employees, no less—expressed disbelief at the claim. Their skepticism matched my own experience; in my years of reporting on Apple and speaking to many of its employees, I had never heard of such a practice. When I sought answers by interviewing current and former Apple engineers, I found that “fake” projects are certainly not a regular occurrence at Apple—and they quite probably do not exist at all.

This article lends credibility to this rumour from 2004:

The device, code-named ‘Q97’ or ‘Asteroid,’ has been under development at the company for the better part of the year. Typically referred to as a ‘breakout box’ in the music recording industry, the external audio device attaches to a computer and offers audio inputs and outputs for attaching instruments or other audio sources. Apple is reportedly building the device around GarageBand, its popular application for aspiring musicians.

Not only did this product seem unlikely, but it was the first time I can remember hearing about “fake” projects inside Apple. Anne Onymus:

What if, rather than creating Asteroid as a real product, the aim of the Asteroid Project was to destroy the rumor sites? All Apple would have to do is assign a few people to developing a semi-viable product be sure to include a “mole” on the team, and have this individual deliberately leak project information to the sites Apple wanted to put out of business.

This always sounded far too conspiratorial to me. Asteroid always seemed like an experimental project, like their current television and smart watch groups.

Johnnie Manzari:

Many designers want to launch a well designed product and have it spread by word of mouth. It feels like the best product should just win. But in situations where the product is facing an incumbent and there are complimentary network effects, it’s simply not enough to launch a well designed product.

The hive effect — especially for social networks — is a particularly strong reason why it’s hard for a newcomer to gain traction. Take Google+: it’s fairly nice to look at, and it’s no more privacy-intrusive than Facebook. But nobody uses it because nobody else uses it (claims to the contrary seem to be based on poor a data collection methodology). Great points from Manzari on how to overcome these effects, at one point referencing an article by Chris Dixon:

A huge challenge for user-generated websites is overcoming the chicken-and-egg problem: attracting users and contributors when you are starting with zero content. One way to approach this challenge is to use what Geoffrey Moore calls the bowling pin strategy: find a niche where the chicken-and-egg problem is more easily overcome and then find ways to hop from that niche to other niches and eventually to the broader market.

I think Google tried to do this by making Google+ invite-only to begin with. But, while that created initial demand through exclusivity, it ultimately restricted the ability for it to be useful. Facebook was successful in a slow rollout because you didn’t have your friends on a centralized social network at the time, whereas Google+ started at a time when your friends were already on Facebook.

Mark Ryan Sallee:

Again, it is crucial to understand that the point of the article was not to test the ability of the car to make the trip. There is no question that if the New York Times author charged the car completely that the trip would pose no problem for the Model S. But charging the car completely takes a lot of time. The goal of the Superchargers is to save time and make electric cars more practical. In my opinion, the charging decisions of the author were practical and reasonable.

I don’t believe Elon Musk is acting intentionally deceptive. His car failed a test, and his natural response was to jump to the defense of the Model S. I do believe Elon’s counter-offense is misguided. It misunderstands what John Broder was aiming to do. And, more worrying, it misunderstands users.

This is one of the best summaries I’ve read of this debacle.

Elon Musk responds to John Broder’s test drive of the Tesla Model S:

When I first heard about what could at best be described as irregularities in Broder’s behavior during the test drive, I called to apologize for any inconvenience that he may have suffered and sought to put my concerns to rest, hoping that he had simply made honest mistakes. That was not the case. […]

When the facts didn’t suit his opinion, he simply changed the facts. Our request of The New York Times is simple and fair: please investigate this article and determine the truth. You are a news organization where that principle is of paramount importance and what is at stake for sustainable transport is simply too important to the world to ignore.

Indeed, this is what I assumed. While the Times is no stranger to controversies with their writers, they are still one of the most reliable, well-regarded news sources in the world. Musk’s data appears to tell a completely different story to Broder’s article, and that has big implications.

If it turns out that Broder faked his story, that means that his journalistic credibility is ruined, and the overall regard of the Times takes another hit. But even if there’s a reasonable explanation for both sides’ accusations, it has proved that the clout of influence has shifted. There could be a bug that displays different information to the driver than what it should. There might be an issue with the graphs Musk presented. But this doesn’t look good for Broder’s article.

Obviously, the Times is going to investigate this. Margaret Sullivan, public editor of the paper:

On #Tesla: I’m on it, as they say. May take some time. Meanwhile, look for a point-by-point response on Wheels blog soon.

Both sides have obvious reasons to embellish the truth, but on the face of it, the data suggests that Broder simply lied. The final graph of Broder trying to run the Tesla flat appears to be damning.

However, Rebecca Greenfield of The Atlantic notes that several of Elon Musk’s claims are not supported by the data. Take the accusation of driving around in circles, apparently trying to run the battery flat:

Musk is accusing Broder of driving three times the most direct route. Maybe Broder missed the charging station and drove around the McDonald’s a couple times looking for it? At the speeds shown in the logs, Tesla says Broder spent around five minutes driving around the service plaza before stopping. If he was deliberately trying to drain the battery, he did not stick to the endeavor for very long.

There’s a lot left to this story.

Update: John Broder delivers his point-by-point rebuttal of Elon Musk’s post. In addition, the Times‘ public editor Margaret Sullivan has delivered her initial impressions on the charges. It seems Mr. Musk is doing damage control more than the Times is.

Matthew Lynley, of the Wall Street Journal‘s Digits blog.

Microsoft is keeping its biggest profit engine tied to its own platforms, Windows 8 and Windows Phone. The company has an incentive to do this: The exclusivity ensures that it’s able to keep selling licenses of Windows 8 and Windows Phone devices.

The problem is that the PC market is starting to shrink in favor of tablets and smartphones, and both of those segments are currently dominated by Apple and Google. Microsoft is doing its best to get in with both Windows Phones and the Surface tablet, but if that doesn’t work it’s running the risk of sacrificing the growth of Office by keeping the umbilical cord with Windows intact.

Scott Hanselman, upon seeing Quora’s requirement to download their app to view more than one answer in a thread:

This implementation goes against everything on the web. You’re not just actively preventing me from visiting your site by forcing me to log in, but you’re also actively forcing me to download your app to access your server.

Quora’s particularly horrible for this. If you sign up for an account, you must select three topics to follow, and then three areas of interest within each topic. I tried to get around that by unfollowing those after completing my registration, but Quora recognized that and required me to maintain those nine followed topics at all times. Quora is incredibly user-hostile.

This time from Tero Kuittinen of BGR:

Around 1994, most mobile phone brands with global ambitions vanished or pulled back to their home markets. The same happened around 2003. We are now drawing close to another mobile Götterdämmerung, a period of destruction and mayhem — and few vendors will survive it.

Around 1994, I was about three feet tall. Around 2003, I was about five-foot-ten. We are now drawing close to another measuring point — by 2014, I will be close to ten feet tall.

Eric Schmidt, October 2010:

“Google policy is to get right up to the creepy line and not cross it,”

Australian app developer Dan Nolan, February 2013:

Let me make this crystal clear, every App purchase you make on Google Play gives the developer your name, suburb and email address with no indication that this information is actually being transferred. With the information I have available to me through the checkout portal I could track down and harass users who left negative reviews or refunded the app purchase.

Seems like an easy fix — don’t record app purchases the same way physical purchases are made — but that’s a lot of very personal information that’s floating around. Google is already requiring people to use their real name when reviewing apps. Via John Gruber.

Jason Fried:

To take it further, maximization as a concept just isn’t interesting to me. I don’t care about maximization. Not maximization of profit, revenue, people, reach, productivity, etc. Not interesting.

Maximization is like kids trying to stay in the treehouse past when their mother calls them in for a snack: that’s when things go wrong. We could do with a lot more people with Fried’s perspective.

Daavid Kahn buys a Nexus 4:

“Ricky” responded: “Thank you for contacting Google Play about the status of your order. I do apologize that you have not received your order. I could see how frustrating that could be. I will do everything I can to assist you.” He assured me that he was conducting a thorough investigation. His second response followed a day later: “Thank you for contacting Google Play about the status of your order. I do apologize for the confusion. I could see how frustrating that could be. I will do everything I can to assist you.” In that moment I was certain – Google had developed a powerful customer support AI, possibly codenamed “HALp”, and I was one of the first customers to interact with him/her/it. So far, it was failing the Turing test.

John Broder stands by his test of Tesla’s Model S sedan.

The problem with all current electric cars — Tesla’s Roadster and Model S, Nissan’s Leaf, Toyota’s electric Prius, and so forth — is that the infrastructure of today doesn’t yet support the cars of tomorrow. But without the early adopters of these vehicles, there will be no demand to build the infrastructure required.

The Nissan Leaf, for instance, will be great around a densely-populated city, provided you remember to charge it at night like you already do with your phone. But in a more sparsely-distributed city or in the suburbs, it’s substantially less effective. Range, charging time, and infrastructure support will get better with time, but these cars are being sold today. These criticisms are valid today, despite Elon Musk’s protests.

It’s a big step in the right direction for these vehicles, though.

Marco Arment took the plunge, so you and I don’t have to:

Asprey’s recipe is 2 tablespoons of butter for 17 ounces of coffee, which is more caffeine than I can handle in the morning, so I made roughly a third of it: 6 ounces of coffee (AeroPress-brewed from 9 grams of beans), and about two-thirds of a tablespoon of unsalted butter.

Butter does not dissolve into coffee. It melts and floats on the surface in a thick, oily layer. I kept stirring it and swirling it around so I wasn’t just taking sips of butter off the top, but it was a losing battle.

My general rule for cooking is to add butter, Parmesan cheese (in big flakes), and cream — in that order. Coffee is the exception. The good news is that, based on this, I don’t need to try adding Parmesan.

Sacha Greif presents the ultimate article regarding that flat/detailed digital/analogue user interface paradigm debate. In short, it’s important to consider why designers make the decision between the two. Both realistic textures and completely flat non-textures are subject to the same ill-considered bandwagon use, and designers need to be judicious.

Rene Ritchie of iMore has compiled a selection of the timeline views of fifteen popular Twitter apps (plus three discontinued ones) for the iPhone. There’s a certain amount of overlap, but many of these are clearly distinct — Tweetbot won’t be confused with Twitterrific, and UberSocial is different from Flurry.

Compare that against a selection of seventeen Android Twitter clients from Phil Nickinson of Android Central. I count eight in there that are nearly identical.

Antriksh Yadav speculates on an implementation of widgets in iOS. I think he’s close, but not quite — think something closer to the way Dashboard is presented in Lion.

Yadav’s pull-down gesture for the lock screen is also weird. Using the same gesture as Notification Centre for something different on the lock screen is inconsistent.

Adam Savage, in a piece for Wired:

Any cursory perusal of a fan/maker forum on the web reveals two distinct kinds of projects: the long, meandering, inconsistently updated but impressively detailed effort and the hell-bent-for-leather, tearing-toward-a-deadline build. Solutions to problems of the first type are often methodical and obvious. Solutions for the second type are much more likely to be innovative, elegant, and shockingly simple.

Over the weekend, I had the opportunity to watch “Sound City”. Even when it became de rigueur, the studio refused to buy a Pro Tools setup. There’s a great line in the film where they comment that “when you’re using 24-track tape, you’re committing to that format”. Constraints have the ability to produce some impressive and innovative solutions that likely wouldn’t arise in a comparatively constraint-free environment.