Commenting on a recent Bloomberg article by Mark Gurman, How Apple Plans to Root Out Bugs, Revamp iPhone Software, Michael Tsai references an otherwise insightful tweetstorm by Steven Sinofsky (a former President of the Windows Division at Microsoft), and shares a few critical observations:

A lot of people are pointing to Steven Sinofsky’s comments. He makes some good points about the “broader context,” but I think he’s completely wrong about Apple’s software quality:

In any absolute sense the quality of Mac/iOS + h/w are at quality levels our industry has just not seen before. […] On any absolute scale number of bugs—non-working, data losing, hanging mistakes—in iOS/Mac is far far less today than ever before.

I don’t see how that can be taken seriously. He doesn’t have access to Apple’s bug database, so how would he know? I really doubt that the number of open bugs is lower than in the past, and even if it were there’s no reason to assume that Radar is representative of the actual number of bugs. He later says that the list of bugs is “infinitely long,” so this whole line of argument seems nonsensical. In what way is today’s Mac/iOS quality better in “any absolute sense” than in, say, 2010? He doesn’t say, except that more people are using it: […]

Well, we can look at how many problems an individual user runs into. Is it higher or lower than before? This measure is independent of Apple’s scale. So is the circle of people I hear complaining. Apple’s customer base has doubled many times over, but the number of family members, friends, and customers that I communicate with has not. Now you could argue that maybe we have become exceptionally unlucky and are running into more than our share of issues, but I don’t find that very convincing.

He wants to discount the actual experiences of “many super smart/skilled people” because “the more a product is used the more hyper-sensitive people get to how it works.” What does that even mean? The number of hours in a day hasn’t increased; I don’t think my Mac/iPhone usage has increased much, if at all. Hardly anyone complains to me about the “slightest changes”; I hear about things that flat out don’t work. That’s not being hyper-sensitive.

I fully agree with Michael here. In a piece I wrote in 2015 — The perceived decline in Apple’s software quality — I argued that “this perceived decline in the quality of Apple’s software products (OS X included) is more related to the nature of the flaws/bugs/annoyances, than the sheer number of those. In other words, it’s not that Apple’s software is quantitatively more buggy today than, say, in the Mac OS 8–9 era, but the issues are (or feel) more critical, and that in turn affects the general level of satisfaction of working with the Mac.” 

At the same time, like Michael and unlike Steven, I can’t say I find today’s Apple software to be far far less buggy or problematic than before. Again, I don’t have access to Apple’s bug database either, so my observations are all necessarily empirical and based on an intensive daily experience with different Macs and different OS X versions. Elaborating on my previous remark, that the perceived decline in Apple’s software quality has more to do with the nature and prominence of the bugs rather than their number, I can say that the latest versions of iOS and Mac OS present a series of annoyances (visual glitches, functional issues, things that work intermittently, etc.) that when manifesting, they have enough prominence and impact to give the whole OS an aura of unpredictability and unreliability; in such a way that, even when everything appears to work just fine, I’m often wondering what kind of issue awaits me round the corner.

And you know what’s ironic? That my experience with Mac OS X High Sierra and iOS 11 has been, for now, limited to borrowed devices and hardware. Devices and Macs I haven’t used as extensively and intensely as my main, older hardware (an iPhone 5 on iOS 10, a MacBook Pro on Mac OS X El Capitan) — and despite the limited usage, I’ve had plenty of occasions to notice buggy behaviours. So this is not being hyper-sensitive towards these issues, because my familiarity with the latest Apple software is only superficial, not developed by an increased usage of a Mac or iPad.

To further corroborate my agreement with Tsai against Sinofsky’s “the more a product is used the more hyper-sensitive people get to how it works” argument, I’ll make a different example, taken from another angle.

I’m still using a fair amount of vintage PowerPC Macs and older iOS devices on a daily basis. I’m writing this on a 17-inch PowerBook G4 from 2003/2004, running Mac OS X 10.5.8 Leopard. I also use other Macs running Tiger (10.4) and even Panther (10.3). I’ve been using these Macs and these versions of Mac OS X constantly for years — and in the case of an iBook G3 and the 12-inch PowerBook G4, since their introduction, April 2005 for Tiger, October 2007 for Leopard. While I indeed encountered a few annoying bugs when Tiger and Leopard were in active development, I remember how the most egregious usually disappeared after a minor OS X release (I even remember resolving an issue on one of my Macs by downloading a Combo Update and reinstalling). 

Whether small or a bit more serious, the bugs, then, felt like something transient passing through an otherwise rock-solid environment. In my 10+ years of using these PowerPC Macs running Tiger and Leopard, I’ve never encountered new issues or noticed things I didn’t before, and I’ve had plenty of time to become ‘hyper-sensitive’ to how they work. Sure, the PowerPC platform isn’t in active development anymore, and I’m speaking of machines and systems that are basically crystallised in their most mature state. But still, in all these years of use, with all the first-party and third-party software I’ve thrown at them, I should have been able to encounter bugs I’d previously missed, or trigger unexpected behaviours. 

While I’m certain there are still underlying issues left unsolved in both Tiger and Leopard, in day-to-day general use, nothing prominent shows up on my radar. I turn on this PowerBook, it boots into Mac OS X 10.5.8, I open whatever apps I need for this session, and I feel I’m working in a stable, predictable environment. The only unfortunate thing I notice is that in places the hardware shows its age, or that certain features or services are too new to support this platform, but neither this particular vintage Mac nor its Mac OS X version are at fault. And it’s pretty amazing I’m still being productive with a 14-year old machine.

I use these PowerBooks, iBooks, and Power Macs, and Mail doesn’t quit unexpectedly or corrupts its message archive; the Finder doesn’t hang randomly, making the machine almost completely unresponsive; after leaving these Macs for a while, I don’t find their fans spinning at maximum speed because a couple of rogue processes are using 134% of CPU resources each(!); their Wi-Fi and/or Bluetooth connection doesn’t drop for apparently no reason at random time intervals (and a couple of those Macs even use third-party Bluetooth dongles!); these and other issues I have instead experienced on more modern Intel Macs with Mac OS X 10.9 and later. And these and other issues are prominent enough to impact the user experience and make people feel distrust towards the operating system and the machine.

I’m just an outside observer, with perhaps the vantage point of having been using Apple hardware for almost 30 years. I can’t say with certainty that today both Mac OS and iOS have more bugs and issues than before. I’m also not saying that everything was 100% perfect before and now it’s all rubbish, because it’s not true. But from having extensively used (almost) each version of Mac OS and iOS, what I do notice is that behind the scenes there was a different approach to their development before a certain point in Mac OS X’s timeline, and that something changed (for the worse) after that point. I roughly place that point between Mac OS X 10.6 Snow Leopard and 10.7 Lion. With iOS things are less clear-cut, because I feel it has always had a lot of attention inside Apple, but the ousting of Scott Forstall clearly was a definite turning point, again not for the better^[1].

Back to Gurman’s article, which originated the whole discussion, on the one hand I really hope that whatever internal software development rethinking process Apple plans to carry out is geared towards recovering part of the old approach to development and quality control I mentioned above; on the other hand, I’m not holding my breath. Not for lack of trust, but because these changes take time and a certain resilience to internal and external pressures.