As a former Apple employee that left in part due to declining software quality (back in 2015!), and the relentless focus on big flashy features for the next yearly release cycle, I could not agree more.
I recently had to do a full reinstall of macOS on my Mac Studio due to some intermittent networking issue that, for the life of me, I could not pin down. Post-reinstall, everything's fine.
I've explained in another thread how this kind of thing happens. It may be the same at other large companies.
Bugs come in (via Radar) and are routed to the team responsible. Ever since Jobs came back (and Apple became valuable again) it has also become very much top-down with the engineers, for better or worse, not calling the shots.
Just an obvious example — there are of course no engineers in the decision to make a "Snow Leopard" release or not. That is a "marketing" decision (well, probably Federighi). But further, even for an engineering team, they're probably not going to be able to make that decision even for their own component(s) either. Again, marketing.
So meetings are held and as it gets close to time to think about the NMOS (next major OS) the team is told what features they will implement. Do you think fix bugs is a feature? How about pay down technical debt? Nope, never.
Fixing bugs is just expected, like breathing I guess. And technical debt ... do what you can given your workload and deliverables. Trust me, many engineers (perhaps especially the older ones) want to both fix bugs and refactor code to get rid of technical debt. But there is simply not the cycles to do so.
And then what is even more insipid, the day the OS ships, every single bug in Radar still assigned to a team, still in Analyze, becomes a much much harder sell for the next OS. Because, you know, you already shipped with it ... must not be that bad.
I'd love to see a bug-fix-only Mac OS release. But I suspect that every time the possibility has come up, something like, I don't know, LLMs burst on the scene and there's a scramble.
> Ever since Jobs came back (and Apple became valuable again) it has also become very much top-down with the engineers, for better or worse, not calling the shots. Just an obvious example — there are of course no engineers in the decision to make a "Snow Leopard" release or not.
It's unclear how much explanatory value this has, because the Snow Leopard that everyone is pining for was during the Jobs era. After all, an Apple that goes bankrupt and out of business isn't going to make any software updates.
I find a stark difference between the Jobs era and the Cook era. Under Jobs, the early Mac OS X updates (Puma and Jaguar) came fast and furious, but then the schedule slowed considerably. Panther was 14 months, Tiger 18, Leopard 30 (delayed due to iPhone), Snow Leopard 22 months, Lion 23. Mountain Lion was the first release after the death of Jobs and came only 12 months after Lion. Thereafter, every Mac OS update came yearly, give or take a few months. That's a drastic change in release schedule.
Yeah, I should be careful to not make it appear as though there were so clear a delineation when Jobs returned. His software engineering team got to work reshaping MacOS (as we know it now) but he seemed to this software engineer to be focused on hardware and "strategies" initially.
Aqua, the new UI, came down from above soon enough. Drawers, toolbars were new UI elements that arrived. In time Jobs' designers were going through the shipping apps with these new UI elements with changes for the engineers to implement.
Certainly by the time the iPhone had arrived the transition to marketing (and design) calling the shots was complete.
Apropos Drawers: The may have looked a little bit silly back then but today almost every Mac app main windows has a big grey sidebar, so that in Exposé view almost all windows look the same. Drawers got an unfair rap, I think.
It's crazy that marketing hasn't worked out that quality and reliability can be spun as a feature. In fact, I remember with OS X, that was the baseline word-of-mouth feature when the comparison was made with Windows at the time.
> Just an obvious example — there are of course no engineers in the decision to make a "Snow Leopard" release or not. That is a "marketing" decision.
I think it is more that the decision to SAY Snow Leopard was a bug fix-only release was a marketing one. The reality is that release also sported things like 64-bit Intel ports of all apps, added Grand Central Dispatch (e.g. an entirely new code concurrency system) and included a from-scratch Finder rewrite.
I always saw these releases (I bundle Mountain Lion in) were all about trying to rein in excessively long release cycles. Short release cycles tend to not have enough time to introduce new bugs, while extended release cycles create a sense of urgency to get code in under the wire.
Now, release cycles have moved to be staged across a fairly predictable annual calendar. If there's an issue where features are getting pushed out 6 months or a year earlier than they should, that is a management and incentives problem.
I don't even know what these big flashy features are anymore. Every year I get asked by staff "Can I upgrade to <latest major Mac OS>" and every time I tell them they can, but they won't see anything different. There's not even big architectural changes under the hood to improve stability or performance.
Short of it being a requirement to use the latest version of Xcode (once they bump the minimum in the following Feburary), and security updates stopping, there's been very little reason to actually upgrade.
>As a former Apple employee that left in part due to declining software quality (back in 2015!), and the relentless focus on big flashy features for the next yearly release cycle, I could not agree more.
Oh Thank You so much. 2013 I was already questioning on some of the features it keeps adding that were useless. Yosemite with continuity was the only useful feature in the past 10 years.
Yes. relentless focus on big flashy features for the next yearly release cycle was exactly what I felt like it was. And that was the big reason why I dislike Craig Federighi.
Edit: Thinking about it more, former Apple employee that worked during 2005 - 2010 is probably a lot more prestige than post 2015.
I recently had to do a full reinstall of macOS on my Mac Studio due to some intermittent networking issue that, for the life of me, I could not pin down. Post-reinstall, everything's fine.