cursor:pointer on buttons

27 points by two 14 hours ago on lobsters | 12 comments

Doesn’t it feel a bit ridiculous that UX design like the material design movement have been fervently removing affordances that tell you “this is an interactible element” for decades making it almost impossible to understand what is a button without actively pointing at every part of the UI, and the solution is to make the pointer more obvious. We shouldn’t have to point at things to understand that they can be clicked.

motet-a | 10 hours ago

Moreover, with a touchscreen and without a mouse, there is no cursor at all. Buttons should definitely look interactible from the first glance.

singpolyma | 3 hours ago

You mean pixel scanning isn't the ideal UI? ;)

dbushell | 11 hours ago

It would be less of an issue if people designed links and buttons to look like links and buttons. The web started with strong blue underlined links and has gone backwards.

Same with modern operating systems (especially from Apple), it's guesswork to know what is interactive. There's no distinctive style, no hover state, not even a label half the time. Just a vague collection of shapes that are suppose to be an icon.

chrismorgan | 12 hours ago

The linked https://kizu.dev/cursor-pointer/ is kinda ridiculous. It dismisses “Don’t break users’ habits!” outright:

But you can’t treat the presence of habits as an argument.

All operating systems have always done it one way. The author is explicitly refusing to consider this to be an argument—not even whether it’s a good or a bad argument, just rejecting its relevance. Ludicrous.

Truth is, there used to be strong customs about button and form control styles, so that interactivity of an element was rarely in question. That has been diminished a lot, and I think that’s why some started adding cursor: pointer to their buttons.

That thesis in turn indicates, in my opinion, why this would be patently a bad decision if it was applied to the base style: the base style of buttons matches the operating system to a reasonable degree (though this is less true than it used to be), so the base cursor should match the operating system’s.

But this is only applied where you’re explicitly changing opting out of that to at least some extent with appearance: base. In that context, I say… meh. Fine. Not enthusiastic about it, but it’s not as unreasonable as it initially seems.

AngryAnt | 12 hours ago

I don't feel strongly in either direction, but did feel that the author does their case a disservice with that dismissal and that the example apps they list off have a distinct web app lean.

Neither would stand out quite so much if title & point was something along the lines of "let's get pragmatic - the battle for halfway consistent interaction design has been lost. Here is our solution for living that reality". I don't know if that is the authors implicit context, but if so it makes a stronger case.

I also wonder if a big part of why their take may be controversial at all is that the usable alternative, in the standard set, to the arrow for this use is the pointing hand. Say the standard set included "inactive arrow" vs. "active arrow" - following the base design language of active vs. inactive buttons - the whole thing would seem like a consistent no-brainer?

polywolf | 10 hours ago

All operating systems have always done it one way. The author is explicitly refusing to consider this to be an argument—not even whether it’s a good or a bad argument, just rejecting its relevance. Ludicrous.

Was this posted in the middle of reading the article?? The very next section addresses the "The cursor doesn't change in the users' OS" argument.

I agree that "don't break users' habits" is too readily dismissed but I don't like to see the rest of the points strawmanned like that. I find the argument "the cursor should change to indicate a click does something" to be more convincing than "buttons vs links has always been done this way"

chrismorgan | 10 hours ago

Those things that followed are called subcategories of that which is to be dismissed.

isuffix | 2 hours ago

Sure it's a subcategory? But they don't dismiss either problem "outright." There's no blanket dismissal, they literally give reasons for why they disagree with the problems. That's just a normal argument.

For the general case of breaking user habits, they say that:

The habit should be treated in its context, and in context of what would happen if we’d break it. Would it be destructive in some way, or would it be just a matter of minor users’ discomfort?

That is a good general reason why you might want to break user habits, and it makes sense to me in this case. I don't think changing the default cursor behavior on the web is likely to affect operating system savants much, and I expect it would help most web users quite a bit!

Then underneath the header "The cursor doesn’t change in users’ OS,"_ they ask whether the OS default is usable, and whether it could be improved. Instead of dismissing that point they claim that some native desktop apps are not usable because of needing to guess where to click on active elements. I agree that that's a problem. Some apps are carefully designed with good affordances, but plenty are shit at showing what things are clickable and interactable. And while that problem isn't solely due to the lack of cursor changes—plenty of other comments already address how better visual design would help—if the OS norm were that cursors should change to pointers over buttons, I think the problem would be better!

I agree with polywolf that your first comment reflects a quick reactive opinion instead of really reading and considering the linked post.

And to make my own point, not breaking user habits is a bad argument for a system that is as long-lived and as universal as the web. Most students in US schools nowadays are growing up with chromebooks instead of laptops with real operating systems. For those users, operating systems are breaking their habits with outdated default behaviors.

[OP] two | 10 hours ago

All operating systems have always done it one way.

atleast android does pointing-finger-cursor on buttons. better to check what the most popular OS in the world does, before typing "All" in emphasis :p

goldstein | 10 hours ago

most android users don’t ever see the cursor, so I’m not sure if “the most popular OS” claim is relevant here

da_rob | 7 hours ago

I do support the general idea, but I guess this never worked for mainstream operating systems because there are so many possible inconsistencies:

What about menu bars? Not a button, so no pointer cursor. Okay. What about the individual items in the menu bar? Not a button, no pointer cursor, aight. So, what about dropdown menus? Needs a pointer cursor, right. So what about the menu items this opens??

What about scrollbar thumbs? No a button, so no pointer cursor, right. But what about the scrollbar buttons??

What about tabbed views, collapsible panels, range sliders, spinner buttons, ...?

When I started working on the Saudade GUI toolkit, I played around with the idea but quickly gave up because it felt so weird and out of place. Maybe I should look at this again.