> should be incremental and security or bug fixes only
> should never update a user interface without allowing the previous one to be used as the default
> should always be backwards compatible with previous plug-ins or other third party add ons
Even as a longtime user of Firefox, who lost access to a number of XUL addons once-upon-a-time, I have to disagree.
Some software upgrades cannot be "incremental or bug-fixes only." Sometimes you get handed a pile of hot garbage, and your best chance is to just refactor it. Certainly, you're not aiming to remove features or break workflows, but there exists a world in which some percentage N of users who are affected negatively by a change are not worth the effort of maintaining something. I think this viewpoint ignores that often as software gets more use and grows, our understanding of the problem-space and relevant abstraction grows as well. Black & white statements about what is or isn't allowed by developers in their products regarding updates completely ignores this aspect of development. Granted, if you can choose when to upgrade and to what version to upgrade this is less of an issue. Modern app stores do not function for the user in this respect.
As for the interface being updated: it happens. Sometimes it is worse, but then many times people only complain about things being "worse" because of assimilation bias. There's some fantastic reading about not necessarily catering to power users: https://pdfs.semanticscholar.org/c011/4ca106ff22cf53aaed57d9.... I don't think I would 100% agree with that full article, but a lot of the central premises hold up.
As for killing compatibility with plug-ins or third party add-ons, I am upset by this as anyone but sometimes it needs to happen. Sometimes a breaking bug is abused by a third party for bad reasons as much as they can be for good reasons (e.g. banning add-ons that use exploits to mine bitcoin or data from users, even if the add-on was offering up a useful feature as well).
Now, relating to Firefox's recent Android update that breaks a lot of things, and even their behaviour for deprecating XUL in the past (I'm not referring to how long they gave people to upgrade, which was fair, but rather that some XUL features were never duplicated in webextensions thereby rendering some extensions impossible to reproduce), I would argue that yeah, maybe there is some sense of truth or some imperative that we need heed in the article. However, I think most of this would be mitigated by being able to pin versions, ask for a specific version from a specific date, or rollback. So maybe it's a wash.
> should never update a user interface without allowing the previous one to be used as the default
> should always be backwards compatible with previous plug-ins or other third party add ons
Even as a longtime user of Firefox, who lost access to a number of XUL addons once-upon-a-time, I have to disagree.
Some software upgrades cannot be "incremental or bug-fixes only." Sometimes you get handed a pile of hot garbage, and your best chance is to just refactor it. Certainly, you're not aiming to remove features or break workflows, but there exists a world in which some percentage N of users who are affected negatively by a change are not worth the effort of maintaining something. I think this viewpoint ignores that often as software gets more use and grows, our understanding of the problem-space and relevant abstraction grows as well. Black & white statements about what is or isn't allowed by developers in their products regarding updates completely ignores this aspect of development. Granted, if you can choose when to upgrade and to what version to upgrade this is less of an issue. Modern app stores do not function for the user in this respect.
As for the interface being updated: it happens. Sometimes it is worse, but then many times people only complain about things being "worse" because of assimilation bias. There's some fantastic reading about not necessarily catering to power users: https://pdfs.semanticscholar.org/c011/4ca106ff22cf53aaed57d9.... I don't think I would 100% agree with that full article, but a lot of the central premises hold up.
As for killing compatibility with plug-ins or third party add-ons, I am upset by this as anyone but sometimes it needs to happen. Sometimes a breaking bug is abused by a third party for bad reasons as much as they can be for good reasons (e.g. banning add-ons that use exploits to mine bitcoin or data from users, even if the add-on was offering up a useful feature as well).
Now, relating to Firefox's recent Android update that breaks a lot of things, and even their behaviour for deprecating XUL in the past (I'm not referring to how long they gave people to upgrade, which was fair, but rather that some XUL features were never duplicated in webextensions thereby rendering some extensions impossible to reproduce), I would argue that yeah, maybe there is some sense of truth or some imperative that we need heed in the article. However, I think most of this would be mitigated by being able to pin versions, ask for a specific version from a specific date, or rollback. So maybe it's a wash.