[Nycto]

Users on mobile can’t make their screens that big, and the average users on desktops don’t expect a webpage to work when they resize their browsers that small. Trying to make both work with the same design can have a negative impact on your customers, from both a usability perspective and by increasing page weight.

Instead, my advice is to create individual designs for each, share when it makes sense, but actively diverge when it’s good for your customers. There doesn’t need to be a single version of a page.

Your mileage may vary.

[hombre_fatal]

There are certainly trade-offs to consider which is one of the reasons UX design demands so much expertise no matter how many HNers dismiss it as some trivial practice (that they, conveniently, never want to do).

Maintaining N versions of your application has costs that aren't necessarily great for your customers either. In my experience it usually cashes out into one version (either mobile or desktop) getting all the support and features while they slowly drip down into the other versions. Meanwhile a responsive design can have the upside of forcing support and feature rollout for all devices simultaneously.

None of this is easy.

[bigDinosaur]

People tend to dismiss UX 'experts' because frequently they end up being the ones who destroy perfectly good interfaces based on trends or similar. The principled ones who adopt a scientific approach are much rarer.

[bluefirebrand]

I don't think it is an issue of "scientific" versus "unscientific"

It's really an issue that many UX designers don't know how browser rendering works, so they design static pages as if they were printing in a magazine.

Pixel perfect mocks are terrible for designing responsive UIs. Trying to build pixel perfect pages in a browser is impossible. Somehow these designers get through school with zero understanding that designing for web is different from designing for print.

[toyg]

It's 2023, I don't think that's really the case any more. If anything, designing for print is now the part of the discipline that has to suffer through "web-isms".

The reality is just that designers gonna design - and designing is often an unscientific craft, pursuing aesthetic values before practical considerations. Google and Apple designers are well-paid and experienced web-heads, and still they led us into a land of well-padded desperation.

[bluefirebrand]

It's 2023 and I am still explaining how responsive design works to new design grads that my company just hired so there's definitely some failure somewhere.

[et-al]

A great recent example is the Slack redesign that didn't improve any user flows and lowers the information density. And that activity badge with sticky notifications. (Shift + ESC is your friend here.)

[withinboredom]

You can’t even exit a search. And god forbid you click a notification while doing a search and have to click “back” 500 times to get to a useable interface.

[znpy]

> People tend to dismiss UX 'experts' because frequently they end up being the ones who destroy perfectly good interfaces based on trends or similar.

yep. i guess that reddit hired one of those.

old.reddit.com is awesome and stood the test of time, the new reddit is awful and slow (and i hate it).

[pbhjpbhj]

There's two things here.

One, not all Reddit users prefer old.reddit (I do).

Two, Reddit aren't designing for users, they're designing for advertiser's to push adverts at users.

Wrt the second point, this means designers aren't designing to the brief you would give them. Like when engineers design obsolescence into a product (it's purposefully inferior for the end user).

Any idiot can see it's bad user experience to keep forcing a user to a design they don't like, but it's not for UX reasons that they do it. The trick is keeping UX good enough.

[TeMPOraL]

> Wrt the second point, this means designers aren't designing to the brief you would give them.

To extend that: note that the very companies that spend most money on UI, that hire the experts and pay them well, that set the trends for entire field of UX - are all companies whose primary business is user abuse - advertising, high engagement, etc. That's what they pay the UI/UX experts to optimize for, and that's what ends up leaking into the wider field - leading astray people who are trying to build things beneficial to their users/customers.

[marginalia_nu]

The ones who adopt a scientific approach are by far the worst. Design is ultimately all about how things ought to be, an act of judgement, meanwhile science is wholly unsuitable for such questions, since it only tells us what is, which following Hume, cannot on its own lead to conclusions about what ought to be.

You get a sort of garbage-in-garbage-out effect if you apply science to a field like design, where it only serves to amplify your own convictions, as what is being fed into the scientific process as unquestioned assumptions inevitably fall out of it as conclusions.

At best you get KPI driven design, which is a vehicle for enshittification, not for building great design.

[hotnfresh]

I'm not following. Science gives us good guidance on what works well or will let us achieve our goals, all the time. It's basically the whole point of doing it at all.

I took the poster as meaning UX that considers the results of, and perhaps even performs, actual user testing & observation, to decide what works and what doesn't. Like operating system vendors used to. I'll grant that "scientific" UX that's just incompetent (99% of the time) application of "telemetry" and A/B testing is awful. But that—and the other bad kind that's just trend-following, personal preference, and whatever will get the best reaction in a design presentation meeting full of non-experts—aren't what I understood as being advocated.

The good kind performs & pays attention to science.

[marginalia_nu]

What science doesn't give good guidance is how to select those goals in a vacuum. The goals at best end up being a version of someone's personal opinion, since there that's the only form of opinion we have access to.

Any opinions you get out of the scientific method were put in there by the person designing the experiment.

[bigDinosaur]

The core of science is iteration based on experiment and hypothesis generation. The scientific approach to UX is simply a hypothesis that a UX change is superior, and an experiment with a measure as to whether or not this is true. Yes, of course you can optimise scientifically for user-antagonistic KPI's (no need to invoke Hume, his point is true and also useless) but the alternative to not adopting a scientific approach is literally just opinion which on the whole is far worse (would you want your airline cockpit's UX to be designed based on a designer's opinion, or real scientific approaches to how people process information and intuit controls?). Of course one would take a good designer's opinion over some MBA-ridden process but a good designer is likely intuiting what would be validated by a more scientific approach anyway.

[itishappy]

Can you relate that to specifics from the article? To my (admittedly non-designer) eyes it appears to be a great example of how science can be used to improve design, and I happen to agree with the findings presented, so I'm curious where you see this breaking down.

[marginalia_nu]

The tacit assumption being made in the article is that good design shouldn't frustrate the users.

I don't disagree with this, but it's none the less an assumption that went into the study, and likewise a conclusion that fell out of it.

[bigDinosaur]

Nobody here is disputing we have values that are not scientific.

[jfim]

> The ones who adopt a scientific approach are by far the worst

I strongly disagree.

Design without considering all of the HCI research that has been done is what you call "garbage-in-garbage-out." We already know how humans perceive information, what makes things salient or invisible, and so on, yet the current design trends completely disregard that with flat UIs and trendy designs that have poor usability.

> At best you get KPI driven design, which is a vehicle for enshittification, not for building great design.

No, you just get trendy design, not usable design.

[andrewprock]

You're more than correct.

In practice, working in small and medium organizations, I have met very few UX designers. Instead I have met plenty of graphical designers that know almost nothing about UX design. I've been at places where I - as a backend developer - know more about practical UX design than anyone on the design team.

I think the reason why we have "bad mobile first design with awful desktop UX" is because very few of the people designing these experiences are UX designers.

I was surprised the article didn't highlight the horror show that is Vector22 at Wikipedia, a design so colossally bad that after three years of suck costs the only path to saving face was to make it the default theme for all users: "Mission Accomplished!"

https://en.wikipedia.org/wiki/Wikipedia:Vector_2022

[rerdavies]

Not getting it. They increased readability, by limiting maximum line length! That is a colossally good thing. Surely, that's like a graphic design 101 kind of decision. (It's a design rule that significantly predates "UX").

The issue at hand is that overly long lines reduce reading speed and comprehension of the content[1]. The optimum length for a digital line of text is somewhere between 66 characters per line and 100 characters per line. I personally use the 100cpl rule. For reference, this HN page has ~185 characters per line on my 1920x1080 display at default scaling.

I do actually remember un-minimizing my browser in order to improve the rate at which I could read the text of ur-Wiki pages.

And then they provided an escape for old men shaking their fists at the sky. Given a choice, I would, without hesitation, choose the new design.

[1] https://en.wikipedia.org/wiki/Line_length#:~:text=characters...

[nicoburns]

Agreed. I had the pleasure to work with talented specialist UX designers early in my careers, and their designs were really fantastic. They also worked really closely with the developers, both to understand the medium they were designing for and as a first line of feedback before things got to real users/clients.

Unfortunately some of the designers I've worked with more recently were primarily graphic designers without a UX background, and actually became an impediment to good design because they were given authority over it despite not really know what they were doing.

I think it's probably an unfortunate consequence of there being more demand for UX designers then there are good UX designers, and simultaneously being a lack of jobs available for graphic designers. And a lot of hiring companies not really understanding what makes a good UX designer.

[pmontra]

I read that page, then browsed Wikipedia a little on desktop. It's a site that I use very often and I didn't notice anything weird. I could have sworn that it has been the theme of Wikipedia for at least 10 years.

I also checked if I had created some rules for that site in Stylus and uBlock Origin, nothing. For once I'm lucky that a change didn't destroy one of my workflows. One could say that if I didn't notice the transition they could have spared themselves all the work, or one could argue that they performed a perfect job.

Anyway, I get directly to the page I need from Google. I found several threads on Reddit complaining about the change and this one https://www.reddit.com/r/wikipedia/comments/10g2cir/im_prett... I see a different usage pattern "all I had to do was open the site and use the search bar. And then from there it was easy to get to the main page, current events, etc." The home page, current events? I'm sure I never heard about current events before now and about the home page, I know that there is one but the search bar of my browser is closer to Wikipedia's internal pages.

[andrewprock]

Without going into details about Vector22, it's certainly better today than it was at launch. It still has a very poor floating ToC UX.

[robertoandred]

How is Vector2022 bad? Text columns that are 100 words wide is bad desktop UX.

[__xor_eax_eax]

Huge engineering cost though. If most of your customers are on mobile, makes sense to optimize for mobile, and hope its "good enough" on desktop.

At the end of the day its all an ROI problem (as are most things)

[bluGill]

The cost need not be huge. Most of the costs should be content, and just the theme is different. However even ignoring that, two themes can be hard if you do them independently. However often only a few changes to one theme are needed to become acceptable, and that is good enough. This in turns means you can limit costs: spend 1 week on making a good desktop theme will already make a big difference as you get the worst offenders fixed.

[spookthesunset]

One problem with two separate designs is deciding when to show one vs. the other. This gets especially tricky when people share links. Wikipedia, for example, has two different URL's: one for mobile and one for desktop. How often do you get links to the mobile version instead of the other?

And if you keep the URL the same but serve different output depending on the browser, then you get inconsistent behaviour between two different devices.

Nailing the UX for mobile and desktop is actually pretty damn hard.

[LocalPCGuy]

I think you're trying to make a very nuanced point, and I tend to agree that there are different needs for different viewport sizes. But I think it's important to note the difference between the design and the technology to build it. The technology should, IMO, as much as possible, seamlessly switch between the various layouts when it makes sense from a viewport size perspective. Definitely don't want, IMO, to deliver completely different sites based on device type/site from a technological point of view, we've tried that before and it isn't a good idea.

I also think one of the things good designers do is to take this into account, and make pages that are built up of components that work at various sizes, not just scaled up from mobile. In addition, a good designer will setup the page design such that it can scale up and down nicely from one viewport size to another.

So, while I don't 100% agree that you need "individual" designs for each, I do think you need a designer that takes the different viewport sizes into account and provides the appropriate adjustments for each. And developers that are skilled at then building those pages.

[fsflover]

> Users on mobile can’t make their screens that big

I can connect my Librem 5 phone to a screen/keyboard and I get a full desktop.

[olyjohn]

It's not mobile when it's connected to a screen / keyboard is it?

[fsflover]

It's never mobile: It runs a desktop GNU/Linux. Dedicated apps are convergent, i.e., automatically change depending on the screen size: https://puri.sm/posts/converging-on-convergence-pureos-is-co...

[kylebenzle]

I split my laptop screen vertical with usually a browser on either side. Occasionally a web page will render itself as if on mobile because it thinks I'm on a narrower screen.

I agree with comment above that it is very hard to make one website responsive to multiple screen sizes.

[robertoandred]

Except you are on a narrower screen. Sounds like correct functionality to me.

[butlike]

You can request the desktop site on mobile, though

[threatofrain]

From an SSR perspective this seems rather hard. How do you correctly identify the user's device at serve time?

[RadiozRadioz]

People say SSR like it's a new concept, but this was how it worked for a long time.

Guess based on user agent (or other fingerprinting metric of choice), redirect to guessed site, provide user the option to override when the page appears, remember the choice in cookie (or local storage).

Though personally I think you can do a lot with responsive CSS if you try hard enough - that is my preferred option.

[progval]

> redirect to guessed site

I always wondered about that. What's the point of redirecting instead of serving a different template on the same URL?

[RadiozRadioz]

That works, you'd just have to put the setting in a cookie rather than local storage.

[progval]

Why? If you can detect mobile user-agents on the first page load, can't you do it in the next page loads too?

[Almondsetat]

Devices should be truthful in the type of content they request. If your phone somehow tells my website that it’s a tablet or a laptop then you should reconsider the intelligence of who has developed your software

[hn_acker]

> Devices should be truthful in the type of content they request.

I think that websites should assume that devices are being truthful. I should be able to request the desktop view on my phone or request the mobile view on my computer. The former I can do sometimes, the latter I can only do with developer tools (and usually doesn't work because the website detects that I'm on desktop!). Browsers could add a header to switch to the mode in which the website dynamically readjusts based on actual device parameters like window size, but by default I need the view to be what I requested regardless of my window size and device type.

You know how Wikipedia has no table of contents on mobile? I made my browser request the desktop site by default so that I could see the table of contents and don't have to tap to open the article sections. (Unfortunately, Wikipedia changed its desktop view UI by moving the table of contents into a hamburger button. On mobile the desktop view forces me to tap the hamburger button to view a blocking popout of the table of contents, while on desktop the contents are automatically opened in a sidebar.) If Wikipedia had forced a dynamic design on me to restrict me to the mobile view on mobile, then I would've wasted time opening article sections to decide whether I wanted to open them in the first place.

[bluGill]

My phone often does tell remote sites it is a desktop because as bad as the desktop experience is, often that is the only way to get at something. (I don't want an app for my doctors office - I check it after my yearly physical and the rest of the time it takes up space I could use for another picture)

[Almondsetat]

Does it do that automatically or does it have an option?

[croes]

If you can't, give me at least a choice