I don't doubt that there is some bias toward remembering more recent complaints more. However, I don't know if you can pull the "it's anecdotal" card on something like this, because unlike most scientific studies (like on the effectiveness of a medical treatment), personal experience is a crucial part of the definition of "software quality." Also, I think it's reasonable to assume that most software, especially in unjailbroken iOS devices, is deterministic, and thus conclude that the experience of a bug is the result of a software problem. I'm a big iOS user, and I definitely share the experience of a steady decrease in software quality.
I should add that while I use OS X and generally install the upgrades promptly, I spend the vast majority of my time in Emacs and Chrome, and thus probably don't really "use OS X" enough to notice any software quality issues. It seems fine to me.
> I think it's reasonable to assume that most software, especially in unjailbroken iOS devices, is deterministic, and thus conclude that the experience of a bug is the result of a software problem
One major factor people tend to forget is that software can end up executing different code paths on different hardware. In other words—software can be both bug-free on a recent Mac, and yet horribly bug-riddled on an older Mac. This produces a very divergent set of personal experiences where people tend to talk over one-another because everyone is seeing a different part of the elephant.
Absolutely so. That's why I was careful to limit my criticism to unjailbroken iOS devices, both because I have much more personal experience with them, and because the hardware and software combinations are much smaller than with OS X.
But even in the example you provide, the experiences of the person with the old bug-riddled Mac are completely valid, and point to real software quality issues (assuming the new OS X version is officially supported on the hardware). The bug-free experience of another person with a recent Mac does not cancel out or in any way diminish of the person experiencing bugs.
Isn't the converse statement just as valid? The buggy experience of a person with an older Mac doesn't provide any useful information about the code quality of OSX for someone considering buying a new Mac. People using a version of OSX on a Mac it could have shipped with, and people using that version of OSX on a Mac that shipped with something completely different in many ways, are separate groups that really only need to communicate their issues within themselves. Crosstalk between them is mostly mudslinging, rather than useful evaluation.
> The buggy experience of a person with an older Mac doesn't provide any useful information about the code quality of OSX for someone considering buying a new Mac.
It absolutely does, because it shows how quickly you will run into trouble if you don't upgrade your hardware. Macs used to be something that the average user could buy every 4-5 years and it will just keep on running well. These users back then would probably never upgrade, but nowadays with free new OSX versions being in your face every time you launch the mac app store, most non technical users will at some point clicky on the button and run into the trap. Meaning that, unlike with earlier software, users have to watch out what they do more so than just a few simple rules like don't delete anything from the trash if you aren't absolutely sure. There's gotchas here and there, so it's not as safe anymore and thus not as empowering (because before, non technical users could be much more bold and try things out, it was harder to get things into a non working state).
> because before, non technical users could be much more bold and try things out, it was harder to get things into a non working state
I'm still wishing any major OS shipped with effective partitioning between storage of OS/application data and user data by default, such that you could hit the "restore to factory settings" button (Cmd+R on OSX boot) and have a guarantee that the only thing that will be blown away is the OS.
It's kind of getting there via an orthogonal path—defaulting users to saving data to the cloud instead of the disk—but it's still not there all-the-way. If it was, I'd just teach my grandparents the "go back to the way it was" button and have much more calming holidays.
Well, at least OSX so far has had a consistent method of resetting applications by deleting the ~/Library/ files, but I'm sure someone at Apple will manage to screw this up via some new iCloud feature sooner or later.
Yes, bugs affecting only older hardware don't give direct useful information about the current code quality on new Macs. However, it does give useful information about the state of the software engineering department at Apple. It's also still relevant to potential buyers of new Macs, because it gives some useful information about the code quality on that hardware years down the road, after updates.
I should add that while I use OS X and generally install the upgrades promptly, I spend the vast majority of my time in Emacs and Chrome, and thus probably don't really "use OS X" enough to notice any software quality issues. It seems fine to me.