The trouble seems to be that they aren't leaving enough functionality available for extensions to bridge the gap. Dropping features to keep the basic weight light and then leaving it to extensions to provide more specific functionality would seem very much in the traditional Firefox style, but that's exactly what they've messed up since moving to rapid releases.
The push for standards came well after as Microsoft overtook Netscape and the Mozilla organization began to push for an open (not tied to Windows / ActiveX) web.
That seems a little one-sided. As I remember it, the biggest battle in the browser war of that generation was probably IE4 against Netscape 4, which was around the time that those really became the dominant browsers and the earlier ones of the Mosaic era finally died out. I'd say it was around the same period that what we might call modern web development was born, with web sites starting to be taken as a more serious form of communication and the arrival of significant numbers of web surfers outside the government and academic communities. It was also around the same period that having actual standards for HTML and CSS started to matter, and during the following few years that Microsoft came in for increasing criticism over their embrace-and-extend strategy in the face of those standards.
That's entirely different than the world that came to exist as Apple, Google and Microsoft are investing heavily in web rendering technology.
There is also the small issue that those three sources represent close to 100% of preinstalled, default browsers today. It's somewhat ironic that just as Microsoft have been paying a bit more respect to web standards with IE10-11, both Google and Apple have overly shunned them. Mozilla have tried to follow suit and certainly haven't done it as well, but they had more to lose in the process. I think that was a huge strategic mistake that will probably lead to the collapse of their business within the next 3-5 years unless something dramatic happens to their product line.
Isn't that what the dev builds are supposed to be for...?
I have no problem with rapidly evolving dev builds. It's using the mainstream user base as permanent beta testers and breaking their experience whether they like it or not that I object to.
It appears our experiences have been almost polar opposites. I have seen few issues with IE10 compatibility in any of the projects I work on, and even fewer for IE11. I have seen so much breakage due to Firefox updates over the past couple of years that it's not even funny, particularly when using anything vaguely new like HTML5 multimedia, canvas/SVG, etc. If I had to sum it up, I'd say Microsoft haven't implemented new features anything like as fast in IE, but when they do implement something that implementation is generally usefully complete and of good quality. Chrome and Firefox implement new stuff all the time, but it can take literally years before the quality is actually acceptable for production use.
I'm not sure exactly what you mean by "browser element" and "web element" or whether your assumptions are therefore realistic. But in any case, how is this any different to any other combinatorial problem for interactions in software architecture? And how come it can't be mitigated by constraining components' behaviour and limiting their ability to interact, also like any other similar problem in software architecture?
We can't know for sure, of course, but the only directly comparable hard data I have shows a huge increase in the number of Firefox compatibility issues we've run into since they went to rapid releases, across a range of different projects using quite different features. Correlation and causation and all that, but this is not a good sign.
The thing with browsers is that you write to a loose standard.
No, the thing with today's browsers is that the "standards" are loose. Up until Google and Mozilla started messing everything up and the W3C became almost completely irrelevant, we enjoyed probably a decade or so where writing web-based front-ends was a big advantage over writing native UIs for each major platform precisely because the web front-end was widely portable with relatively little effort and good reliability.
We're repeating exactly the same mistakes as we did in the IE vs. Netscape embrace-and-extend days, just now with more players and -- perhaps a greater cause for concern -- with most of the leading players having a much stronger bias towards driving the Web in a specific direction that favours their own commercial goals at the expense of other valuable uses for Web technology.
Isolation between tabs is not much of a problem in practice, at least with Firefox
I'm glad it's not a problem for you. Personally, I seem to run into so many crashes, sometimes with plug-ins but sometimes with bugs in FF itself, that it's a serious problem in practice.
This is about far more than just UIs, though the constantly mutating UIs are infuriating to be sure.
Rapid automatic updates mean everyone has the latest version, which means developers can count on everyone having the latest version.
The thing is, I don't want to count on everyone having the latest version. I want to be able to test my site or app, and to know that if it works in testing and I push out to production, my users will enjoy the same fully working system I signed off. And they will still be able to enjoy the same fully working system tomorrow, and next week, and next month.
Bleeding edge features are of little interest to me, because approximately 0% of them will work reliably across all major browsers when they are first introduced, or even across all of the evergreen ones. I'm not using the latest cutting edge ES6 support, I'm transpiling to reliable, portable, stable ES5 with Babel, like almost every other JS developer I know in 2015. I'm not using flexbox and cute animation tricks, because there are too many bugs to make them reliable.
In any case, while some of these tools would have been neat five years ago, today we've already solved many of the real world problems they address. While our solutions might not be as elegant, they are tried and tested, and they already exist. I'm not about to rewrite my more-than-five-minutes old web app, which works just fine for my users already, to incorporate the newly blessed shiny that might work in most browsers if I'm lucky.
Just about every aspect of modern UI counts on this.
No. I'm sorry, but that's just not true. I don't know your background, but as someone who has multiple web-related businesses and does a fair bit of freelance and consultancy work, I would wager that I work on a wider variety of real world web projects than most people reading this. Some of those projects have relatively advanced UIs, and some of them are relatively large and long-lived as web projects go. And I cannot think of a single time that any of those projects has been able to take advantage of some new browser feature that came out within the past six weeks. Not once. Ever.
Many of those projects have suffered significantly due to the ever-changing bug landscape and feature support in evergreen browsers, though. It's a huge drain on developer productivity and customer support.
Try taking IE 8 for a spin sometime, it's awful.
This argument makes no sense. IE8 is also nearly 7 years old. Even if browsers only issued a new stable release every six years, IE8 still wouldn't be the current version. And in my experience, basically no-one in 2015 is still clinging to IE8 outside of perhaps a few very large and very slow-moving businesses in specific industries.
And of course don't get me started on the Security nightmare that happens when you've got dozens of unsupported browser versions in use because people refuse to upgrade.
This is also a fundamentally flawed argument. You're conflating security updates with functionality updates, which is almost never actually necessary. It is perfectly possible to have a stable functional base and UI but apply rapid patches that are essential for security. Ask anyone who runs a Debian server, for example.
Basically you're point is only valid if you ignore the mountains of under the hood enhancements that have been piling into browsers for the last 10 years.
Ten years ago, Firefox was in its infancy, IE6 was state of the art, and Chrome wouldn't exist at all for several more years. You're just making this up now.
The alternative to regular forced upgrades is a complex series of radically different standards.
How do you figure that? The Web evolved from infancy to arguably the most effective communications system and knowledge repository in the history of humanity without needing six-weekly updates. And frankly, developing for the Web was much easier when web standards actually meant something too, while trying to keep up with this week's bug in Chrome or Firefox is a horrendous drain on productivity and morale. For all its flaws, as least you knew where you stood with something like IE6, and once you'd figured out the handful of workarounds you needed if you wanted to use a newer feature, most of the time stuff just worked.
Ultimately product owners focused on a specific product are going to do a better job than the vast majority of users in deciding how to move their products forward.
The trouble is, it's not clear that the Web and browsers generally and Firefox in particular are moving forward. They're moving for sure, but all too many changes in the relatively recent past have been steps backward for significant numbers of users. We could debate specifics, or we could just look at Firefox's market share dropping like a rock as it has steadily eroded the priorities and flexibility that made it an attractive choice for so long.
It's funny how back when Chrome and then Firefox introduced rapid, automatic updates to the web-browsing public, I would get modded to oblivion if I expressed my opinion that this is a bad thing and not in users' interests. It means features keep breaking and UIs keep moving around and sometimes useful functionality even gets removed entirely. Moreover, real web sites and apps don't tend to use bleeding edge features anyway, because those features aren't stable and reliable across browsers, so the main benefit claimed for very rapid updates is mostly an illusion. I suspect a lot more people would agree with me today, though.
I still do not understand why it is so hard to have a flexible UI.
One factor is that the underlying codebase evidently has some significant architectural problems, which unfortunately Mozilla haven't been able to resolve in a long time. This line gets wheeled out time and again to explain why Firefox still doesn't support important features like proper isolation between tabs, and sometimes also for more minor issues like why security warnings sometimes don't match up with what's actually happening on the page.
I can't help thinking that if they had focussed on getting their software architecture house in order first, before all the whizzy new features and never-ending UI rearrangements that no-one actually seems to want, Firefox would look and feel a lot different today. I see happy users citing Pale Moon every time these discussions come up now, and perhaps that's why.
Different rules do apply when a human being does it. Laws certainly vary from one jurisdiction to another, but I don't know any jurisdiction where infringement of intellectual property rights is treated the same in law as theft of a physical item.
What goods were stolen? We're talking about intellectual property here, so legally speaking a different set of rules applies (even if various propaganda from rightsholders likes to pretend it's the same as real property).
There have been a few attempts to apply the basic idea of copyright to lock down derivative works, APIs, and other oblique uses of the original material. If you're not careful you can run into trademark issues as well. These suits tend to rely on dubious legal technicalities, but the trouble with dubious legal technicalities is that they are still what the letter of the law actually says, so a lot of things that seem like they shouldn't be affected by intellectual property laws as originally envisaged can in practice be affected anyway... particularly if you're in a legal system where the guy with the most money has a huge advantage in practice.
The problem of how to effectively enforce the law on the enforcers of the law is a very difficult one.
That's certainly a valid concern, though I think it's a separate one to the reasons lawyers are normally ethically bound to represent any client to the best of their ability in an adversarial system.
The us-and-them culture you refer to also seems to be particularly strong in the US, and it appears from the outside to be a serious and growing problem, though that could just be bias in which news makes it to my part of the world. If you asked most police officers here in the UK, I think you would find that community relations are valued at all ranks, and there is a great deal of genuine concern within the police service itself about political meddling and orders from on high that can and in some cases do jeopardize those community relations, particularly following a series of high-profile screw-ups on that score in recent years.
But a corrupt criminal justice system is SO MUCH WORSE than ANY number of other sorts of bad guys that it's a good trade.
Yes. The fruit of the poisoned tree argument is another of those greater-good principles I was talking about.
The enhanced surveillance that the FISA court has authorized the NSA to conduct is the best way to identify these terrorists before they attack us. We don't need a repeat of the Paris attacks or 9/11, and the surveillance is the single best tool to prevent those things from happening again. Why are you guys so obsessed with making America less safe and making it easier for terrorists to operate?
I'm going to assume this is a sincere question and not just a troll. The two main answers would be because there is precious little evidence that routine mass surveillance does in fact make anyone safer, and because that form of surveillance may have additional much less desirable consequences.
A lot of the basic principles of how our legal systems work are based on greater-good arguments like this. The ethical requirement for a lawyer to represent their client faithfully and to the best of their ability is probably necessary for an adversarial court system to function effectively, for example.
The trouble with greater-good arguments is that someone usually winds up the unlucky one, and if we're talking about legal/government issues, the consequences for the unlucky one can be severe. Whether your the hostage whose ransom wasn't paid, or the innocent who was mistaken for a terrorist by armed police, or the victim in court who has to relive a horrific assault from the witness box under cross-examination by the defence, or the wrongly convicted "murderer" sent to jail for life, it wasn't your fault that you got screwed by the system, but you still got screwed all the same.
Thus we also get ethical arguments that it is better to let 10 guilty men go free than to convict 1 innocent man and so on. That's great if you're the innocent who got the benefit of the doubt, but not so great if you're the next victim of the 10 guilty ones who also got the benefit of the doubt. There are no easy answers to these issues. For some, there are probably no perfect answers at all.
The trouble with fighting for human freedom is that one spends most of one's time defending scoundrels. For it is against scoundrels that oppressive laws are first aimed, and oppression must be stopped at the beginning if it is to be stopped at all.
Swearing at me won't make your argument any more convincing.
Also, extended support for XP only ended in April 2014, more than 12 years after its release to manufacturing, while extended support for Vista still has a year and a half to run despite that one being the turkey and no fewer than four newer versions of desktop Windows being released since then. It's right there on Microsoft's web site.
Why? People keep making these sound-bite replies as if I'm asking for something that is somehow unrealistic, ignoring the fact that what I described has been absolutely routine for multiple decades. And at Enterprise level, it seems even 2015 Microsoft aren't willing to rock that boat.
This idea that it's acceptable to ship sub-standard software and then fix it later (very possibly breaking other stuff in the process) is a relatively recent invention, and it needs to die. It can be buried next to the idea that rapid update cycles make better software from the user's point of view.
I want to be able to run the OS it had when I bought it. That was the OS I chose to pay money for and for whatever reason decided was what I wanted to use.
Sometimes newer versions are better, but sometimes they aren't. Being able to run some new version of an OS on older hardware is nice as an option, but not my primary requirement.
We have some Apple gear for testing, and for the most part we've also upgraded OK. We know other people who weren't so lucky.
Anecdotal data is anecdotal, but my point is that there is some risk involved with any major software upgrade. Users shouldn't be forced to take that risk just to get security patches for vulnerabilities that, obviously, shouldn't have been there in the first place.
Slashdot Mirror
The trouble seems to be that they aren't leaving enough functionality available for extensions to bridge the gap. Dropping features to keep the basic weight light and then leaving it to extensions to provide more specific functionality would seem very much in the traditional Firefox style, but that's exactly what they've messed up since moving to rapid releases.
The push for standards came well after as Microsoft overtook Netscape and the Mozilla organization began to push for an open (not tied to Windows / ActiveX) web.
That seems a little one-sided. As I remember it, the biggest battle in the browser war of that generation was probably IE4 against Netscape 4, which was around the time that those really became the dominant browsers and the earlier ones of the Mosaic era finally died out. I'd say it was around the same period that what we might call modern web development was born, with web sites starting to be taken as a more serious form of communication and the arrival of significant numbers of web surfers outside the government and academic communities. It was also around the same period that having actual standards for HTML and CSS started to matter, and during the following few years that Microsoft came in for increasing criticism over their embrace-and-extend strategy in the face of those standards.
That's entirely different than the world that came to exist as Apple, Google and Microsoft are investing heavily in web rendering technology.
There is also the small issue that those three sources represent close to 100% of preinstalled, default browsers today. It's somewhat ironic that just as Microsoft have been paying a bit more respect to web standards with IE10-11, both Google and Apple have overly shunned them. Mozilla have tried to follow suit and certainly haven't done it as well, but they had more to lose in the process. I think that was a huge strategic mistake that will probably lead to the collapse of their business within the next 3-5 years unless something dramatic happens to their product line.
Isn't that what the dev builds are supposed to be for...?
I have no problem with rapidly evolving dev builds. It's using the mainstream user base as permanent beta testers and breaking their experience whether they like it or not that I object to.
It appears our experiences have been almost polar opposites. I have seen few issues with IE10 compatibility in any of the projects I work on, and even fewer for IE11. I have seen so much breakage due to Firefox updates over the past couple of years that it's not even funny, particularly when using anything vaguely new like HTML5 multimedia, canvas/SVG, etc. If I had to sum it up, I'd say Microsoft haven't implemented new features anything like as fast in IE, but when they do implement something that implementation is generally usefully complete and of good quality. Chrome and Firefox implement new stuff all the time, but it can take literally years before the quality is actually acceptable for production use.
I'm not sure exactly what you mean by "browser element" and "web element" or whether your assumptions are therefore realistic. But in any case, how is this any different to any other combinatorial problem for interactions in software architecture? And how come it can't be mitigated by constraining components' behaviour and limiting their ability to interact, also like any other similar problem in software architecture?
We can't know for sure, of course, but the only directly comparable hard data I have shows a huge increase in the number of Firefox compatibility issues we've run into since they went to rapid releases, across a range of different projects using quite different features. Correlation and causation and all that, but this is not a good sign.
The thing with browsers is that you write to a loose standard.
No, the thing with today's browsers is that the "standards" are loose. Up until Google and Mozilla started messing everything up and the W3C became almost completely irrelevant, we enjoyed probably a decade or so where writing web-based front-ends was a big advantage over writing native UIs for each major platform precisely because the web front-end was widely portable with relatively little effort and good reliability.
We're repeating exactly the same mistakes as we did in the IE vs. Netscape embrace-and-extend days, just now with more players and -- perhaps a greater cause for concern -- with most of the leading players having a much stronger bias towards driving the Web in a specific direction that favours their own commercial goals at the expense of other valuable uses for Web technology.
Isolation between tabs is not much of a problem in practice, at least with Firefox
I'm glad it's not a problem for you. Personally, I seem to run into so many crashes, sometimes with plug-ins but sometimes with bugs in FF itself, that it's a serious problem in practice.
This is about far more than just UIs, though the constantly mutating UIs are infuriating to be sure.
Rapid automatic updates mean everyone has the latest version, which means developers can count on everyone having the latest version.
The thing is, I don't want to count on everyone having the latest version. I want to be able to test my site or app, and to know that if it works in testing and I push out to production, my users will enjoy the same fully working system I signed off. And they will still be able to enjoy the same fully working system tomorrow, and next week, and next month.
Bleeding edge features are of little interest to me, because approximately 0% of them will work reliably across all major browsers when they are first introduced, or even across all of the evergreen ones. I'm not using the latest cutting edge ES6 support, I'm transpiling to reliable, portable, stable ES5 with Babel, like almost every other JS developer I know in 2015. I'm not using flexbox and cute animation tricks, because there are too many bugs to make them reliable.
In any case, while some of these tools would have been neat five years ago, today we've already solved many of the real world problems they address. While our solutions might not be as elegant, they are tried and tested, and they already exist. I'm not about to rewrite my more-than-five-minutes old web app, which works just fine for my users already, to incorporate the newly blessed shiny that might work in most browsers if I'm lucky.
Just about every aspect of modern UI counts on this.
No. I'm sorry, but that's just not true. I don't know your background, but as someone who has multiple web-related businesses and does a fair bit of freelance and consultancy work, I would wager that I work on a wider variety of real world web projects than most people reading this. Some of those projects have relatively advanced UIs, and some of them are relatively large and long-lived as web projects go. And I cannot think of a single time that any of those projects has been able to take advantage of some new browser feature that came out within the past six weeks. Not once. Ever.
Many of those projects have suffered significantly due to the ever-changing bug landscape and feature support in evergreen browsers, though. It's a huge drain on developer productivity and customer support.
Try taking IE 8 for a spin sometime, it's awful.
This argument makes no sense. IE8 is also nearly 7 years old. Even if browsers only issued a new stable release every six years, IE8 still wouldn't be the current version. And in my experience, basically no-one in 2015 is still clinging to IE8 outside of perhaps a few very large and very slow-moving businesses in specific industries.
And of course don't get me started on the Security nightmare that happens when you've got dozens of unsupported browser versions in use because people refuse to upgrade.
This is also a fundamentally flawed argument. You're conflating security updates with functionality updates, which is almost never actually necessary. It is perfectly possible to have a stable functional base and UI but apply rapid patches that are essential for security. Ask anyone who runs a Debian server, for example.
Basically you're point is only valid if you ignore the mountains of under the hood enhancements that have been piling into browsers for the last 10 years.
Ten years ago, Firefox was in its infancy, IE6 was state of the art, and Chrome wouldn't exist at all for several more years. You're just making this up now.
The alternative to regular forced upgrades is a complex series of radically different standards.
How do you figure that? The Web evolved from infancy to arguably the most effective communications system and knowledge repository in the history of humanity without needing six-weekly updates. And frankly, developing for the Web was much easier when web standards actually meant something too, while trying to keep up with this week's bug in Chrome or Firefox is a horrendous drain on productivity and morale. For all its flaws, as least you knew where you stood with something like IE6, and once you'd figured out the handful of workarounds you needed if you wanted to use a newer feature, most of the time stuff just worked.
Ultimately product owners focused on a specific product are going to do a better job than the vast majority of users in deciding how to move their products forward.
The trouble is, it's not clear that the Web and browsers generally and Firefox in particular are moving forward. They're moving for sure, but all too many changes in the relatively recent past have been steps backward for significant numbers of users. We could debate specifics, or we could just look at Firefox's market share dropping like a rock as it has steadily eroded the priorities and flexibility that made it an attractive choice for so long.
It's funny how back when Chrome and then Firefox introduced rapid, automatic updates to the web-browsing public, I would get modded to oblivion if I expressed my opinion that this is a bad thing and not in users' interests. It means features keep breaking and UIs keep moving around and sometimes useful functionality even gets removed entirely. Moreover, real web sites and apps don't tend to use bleeding edge features anyway, because those features aren't stable and reliable across browsers, so the main benefit claimed for very rapid updates is mostly an illusion. I suspect a lot more people would agree with me today, though.
I still do not understand why it is so hard to have a flexible UI.
One factor is that the underlying codebase evidently has some significant architectural problems, which unfortunately Mozilla haven't been able to resolve in a long time. This line gets wheeled out time and again to explain why Firefox still doesn't support important features like proper isolation between tabs, and sometimes also for more minor issues like why security warnings sometimes don't match up with what's actually happening on the page.
I can't help thinking that if they had focussed on getting their software architecture house in order first, before all the whizzy new features and never-ending UI rearrangements that no-one actually seems to want, Firefox would look and feel a lot different today. I see happy users citing Pale Moon every time these discussions come up now, and perhaps that's why.
Different rules do apply when a human being does it. Laws certainly vary from one jurisdiction to another, but I don't know any jurisdiction where infringement of intellectual property rights is treated the same in law as theft of a physical item.
What goods were stolen? We're talking about intellectual property here, so legally speaking a different set of rules applies (even if various propaganda from rightsholders likes to pretend it's the same as real property).
There have been a few attempts to apply the basic idea of copyright to lock down derivative works, APIs, and other oblique uses of the original material. If you're not careful you can run into trademark issues as well. These suits tend to rely on dubious legal technicalities, but the trouble with dubious legal technicalities is that they are still what the letter of the law actually says, so a lot of things that seem like they shouldn't be affected by intellectual property laws as originally envisaged can in practice be affected anyway... particularly if you're in a legal system where the guy with the most money has a huge advantage in practice.
The problem of how to effectively enforce the law on the enforcers of the law is a very difficult one.
That's certainly a valid concern, though I think it's a separate one to the reasons lawyers are normally ethically bound to represent any client to the best of their ability in an adversarial system.
The us-and-them culture you refer to also seems to be particularly strong in the US, and it appears from the outside to be a serious and growing problem, though that could just be bias in which news makes it to my part of the world. If you asked most police officers here in the UK, I think you would find that community relations are valued at all ranks, and there is a great deal of genuine concern within the police service itself about political meddling and orders from on high that can and in some cases do jeopardize those community relations, particularly following a series of high-profile screw-ups on that score in recent years.
But a corrupt criminal justice system is SO MUCH WORSE than ANY number of other sorts of bad guys that it's a good trade.
Yes. The fruit of the poisoned tree argument is another of those greater-good principles I was talking about.
The enhanced surveillance that the FISA court has authorized the NSA to conduct is the best way to identify these terrorists before they attack us. We don't need a repeat of the Paris attacks or 9/11, and the surveillance is the single best tool to prevent those things from happening again. Why are you guys so obsessed with making America less safe and making it easier for terrorists to operate?
I'm going to assume this is a sincere question and not just a troll. The two main answers would be because there is precious little evidence that routine mass surveillance does in fact make anyone safer, and because that form of surveillance may have additional much less desirable consequences.
A lot of the basic principles of how our legal systems work are based on greater-good arguments like this. The ethical requirement for a lawyer to represent their client faithfully and to the best of their ability is probably necessary for an adversarial court system to function effectively, for example.
The trouble with greater-good arguments is that someone usually winds up the unlucky one, and if we're talking about legal/government issues, the consequences for the unlucky one can be severe. Whether your the hostage whose ransom wasn't paid, or the innocent who was mistaken for a terrorist by armed police, or the victim in court who has to relive a horrific assault from the witness box under cross-examination by the defence, or the wrongly convicted "murderer" sent to jail for life, it wasn't your fault that you got screwed by the system, but you still got screwed all the same.
Thus we also get ethical arguments that it is better to let 10 guilty men go free than to convict 1 innocent man and so on. That's great if you're the innocent who got the benefit of the doubt, but not so great if you're the next victim of the 10 guilty ones who also got the benefit of the doubt. There are no easy answers to these issues. For some, there are probably no perfect answers at all.
The trouble with fighting for human freedom is that one spends most of one's time defending scoundrels. For it is against scoundrels that oppressive laws are first aimed, and oppression must be stopped at the beginning if it is to be stopped at all.
-- H. L. Mencken
Swearing at me won't make your argument any more convincing.
Also, extended support for XP only ended in April 2014, more than 12 years after its release to manufacturing, while extended support for Vista still has a year and a half to run despite that one being the turkey and no fewer than four newer versions of desktop Windows being released since then. It's right there on Microsoft's web site.
Why? People keep making these sound-bite replies as if I'm asking for something that is somehow unrealistic, ignoring the fact that what I described has been absolutely routine for multiple decades. And at Enterprise level, it seems even 2015 Microsoft aren't willing to rock that boat.
This idea that it's acceptable to ship sub-standard software and then fix it later (very possibly breaking other stuff in the process) is a relatively recent invention, and it needs to die. It can be buried next to the idea that rapid update cycles make better software from the user's point of view.
I want to be able to run the OS it had when I bought it. That was the OS I chose to pay money for and for whatever reason decided was what I wanted to use.
Sometimes newer versions are better, but sometimes they aren't. Being able to run some new version of an OS on older hardware is nice as an option, but not my primary requirement.
What you think are different OSes are really just security patches and upgrades to the same underlying OS.
Some of those "upgrades" include completely changing the look and feel several times, loss of compatibility with older software...
For practical purposes, each major revision of OS X is typically a new OS. Windows 8 was not Windows 7 just because it had Windows in its name.
Apparently I'm still living in that "software fantasy land". I just run Windows 7 instead of Windows 10.
We have some Apple gear for testing, and for the most part we've also upgraded OK. We know other people who weren't so lucky.
Anecdotal data is anecdotal, but my point is that there is some risk involved with any major software upgrade. Users shouldn't be forced to take that risk just to get security patches for vulnerabilities that, obviously, shouldn't have been there in the first place.