Thursday, 7 May 2015

The Straw That Broke the Camel’s Back

Apple claims to have $4.7 billion goodwill in its Q2 of 2015 report. Roughly a 2% increase on its number for its 2014 fiscal year. I question this given longtime influential Apple users Marco Arment and Craig Hockenberry seemingly harbouring less goodwill towards Apple of late. See their respective pieces titled “Apple has lost the functional high ground” and “Death by a thousand cuts”. Craig Hockenberry said in another post lamenting Apple’s software quality decline “[…] we now have more empathy for the bullshit that folks using Windows have suffered with for years.” This resonated with me. I now feel a sense of powerlessness using current Apple products (Watch included) which I have only ever felt before when using Windows.

A recent discovery I have made with Apple’s Health app—whose release or lack thereof was another fiasco—is significant for no other reason then it is the straw that broke the camel’s back. I can’t resist putting my proverbial pen to paper and adding my thoughts to the Apple quality decline discourse..

I recently started recording my weight using the Health app. My measurements not being spaced regularly, led me to discover a potential misnomer in the interface. I will concede the pettiness of this complaint on its own, however when put into the above context, it betrays the recent, systematic and disappointing lack of scrupulousness on Apple’s part. I am fairly certain that my “daily average” weight is not 92.57 based on the data I entered into Health:

Let $A$ be the set of data points in the health app and $B$ be the set of durations such that $B_{i}$ is the duration between the date of measurement $A_{i}$ and the date of measurement $A_{i + 1}$ . Apple’s “daily average” is simply:

\frac{1}{| A |} \sum_{a \in A} a = 92.57

The correct daily average adjusted for the time between measurements is:

\frac{\sum_{i = 0}^{| B |} B_{i} \frac{A_{i} + A_{i + 1}}{2}}{\sum_{b \in B} b} = 92.12

I copied the numbers into Google Sheets as Numbers doesn’t support time series chart (sigh) to help illustrate the difference. It should be visually clear that the weighted average is closer to the middle of the long downwards slope, and therefore closer to the true overall daily average.

This mistake, albeit relatively small, is one that I would not expect from talented engineers. This has me wondering if Apple’s quality assurance (QA) woes of late stem not from poor QA, but from poor engineering. The above has me concerned that it is the latter, a far more serious issue for Apple as it betrays serious talent retention problem, which could take years to put right.

I won the lottery this year and will be going to WWDC in June. I hope Apple will impress us and turn things around. If not, I may well be attending Google IO in 2016.

This article has been altered since its original release on 7 May 2015:

A paragraph about my recent excellent experience updating my Nexus 4 to Android Lollipop has been removed as it detracted from the main issue.
A conclusion has been added.