Why can’t we just remove user-hostile features like this from browsers? At this point I’m starting to feel like we need to start over. Create a new web, focused on the user, that doesn’t have any of this nonsense.
Which is also hostile. Why not use something like an img/svg?
If you think it shouldn't be possible to develop user interfaces that rival the desktop inside the browser, then just say so. If you merely believe existing web standards make this possible without "hacks" like "user-select:none", then you are mistaken.
By the way, I'm a "read with selection" guy myself and I find this behavior on paragraphs of text unacceptable. On the other hand, none of the text on that website is worth reading anyway.
I gave you several examples of these. The text for many UI elements isn't supposed to be selectable, because it interferes with user interaction.
Your off-chance example where somebody is trying to copy-paste the text for translation because the application is in a foreign language is subordinate to this.
Good UX is about choosing the right tradeoffs between functionality and usability.
It would make sense, and all text in a user interface should be user selectable and offer the same behaviors.
Maybe you need to copy and paste a lengthy error message. Maybe there’s a word you don’t understand you want to look up in a dictionary. Maybe the interface isn’t in your native language and you want to translate it.
The fact that most mainstream computing environments don’t offer this in 2019 doesn’t mean it’s the right thing to do, rather that our systems have regressed in some (many?) areas when it comes to usability.
How would you know? You have never used such a system because it doesn't exist, how do you know it's better?
If you had ever implemented something like a drag and drop UI, a context menu, or even a a simple button using plain old HTML, you would've immediately noticed that selectable text interferes with user interaction. It's the first thing you disable and it's the reason why "user-select:none" must exist.
I have, it does, it's called Oberon, it was glorious.
Copying text from UI elements was also never an issue for me in GTK2 (can't speak for 3, haven't used it much), it was great (although I admit I don't know if there were a few places in the system where I couldn't copy text, it may not have been 100% the case).
When I'm working in terminal apps (more often than you can likely imagine) on Linux I can also expect to copy text everywhere, it's great.
I'd encourage you to use a diversity of computing systems, it'll broaden your perspective and imagination and sense of possibilities :)
I have used text interfaces, but they are inferior for most users and many types of applications.
> I'd encourage you to use a diversity of computing systems, it'll broaden your perspective and imagination and sense of possibilities :)
I suggest the same to you, but at the application level, not at the "computing system" level. There are a lot of possibilities with graphical user interfaces that aren't limited to text.
Usually it's possible by just starting the selection beside the button and dragging across. For example, you can select a navigation button on HN easily this way, along with most sites. I use this all the time when I'm reading about products and want to copy the product name or model number, but it's part of a breadcrumb navigation. It would actually be nice if there was a browser shortcut, where if I hold "s" I can disable opening links during that time and select buttons like normal text.
I wonder about what kind of a performance decrease you'd expect from being able to blacklist or whitelist chunks of the web API, like how you can blacklist syscalls in some container platforms.
Are they run-time prefs that can be toggled? There is UI/UX overhead, and related attack surface.
If they are run-time prefs, are they loaded when the browser is run, or on each page load/content type load? Do you have to maintain a list of prefs by domain? etc, etc, etc.
If they are compile time, then is it even worth doing for dramatic increase in complexity trying to support users with unusual configuration (e.g. THAT user that decides that preffing off JSON support is a good idea.)
How granular do you allow user prefs on this, and how do you account for the fact that a user may pref off an api/feature that you use as a part of exploit mitigation?
Do you want to log entries that are pref'd off so that users know what is making their favorite app fail?
What are the abuse cases for pref'd features?
How much attack surface is introduced by allowing pref'ing of features?
Does doing this actually contribute to usability on the web, or will it simply lead to a proliferation of "Please ensure that all features are pref'd on" pages, similar to the Javascript pages you often see when browsing without javascript.
One would argue that there would be bloat -- having to install multiple extensions. But then visiting random websites is completely safe, and the only ones that we have to care about is the popular ones that many people have extensions for.