Hacker News Reader

Comment by hsn915 | Hacker News Reader

In almost every company I joined, the common approach to design anything is to look at what everyone else is doing and copy it.

"Google does it this way, so we should too".

"Facebook has spent millions on research and decided to do it this way, we should copy them".

If you do anything different, you are immediately questioned:

"Why are you doing it like this? No one else does it this way".

The tech industry is basically controlled by non-tech and non-design people who have no unique sense or taste in anything to speak of. Even self-branded "UX Designers".

View on HN

I’ve enjoyed some UX PhD research seminars, and I think something needs to be said for the legitimate UX research being done. (You wouldn’t believe the thought/effort/testing that can go into the shape, color, details, etc. of a web app button.) For example: Amazon’s app structure (for their main online shopping experience) has been tested ad-infinitum by people who know what they’re doing. If you’re designing a similar shopping experience, it’s not “copying Amazon” as much as using the state of the art in shopping experience design.

There’s no deductive logic that says “Facebook has spent millions on research so it’s correct”, but common sense says this isn’t a bad strategy. UX research is a legitimate undertaking, and Facebook definitely had the chops to hire people who have invested their education in understanding the nuances of when to use which design elements. It’s literally a science.

UX design (I’m not speaking of other aspects of design) isn’t really a creative process as much as an engineering discipline. There’s a best way to achieve any particular outcome. Maybe taste is involved in where you hide the rubbish of your application, but the basic functionality has a “best flow”. Copying one of the giants is the best way to get this flow for the average person.

The Amazon shopping experience was actually heavily dictated by Jeff's personal viewpoints, with the (correctly stated) very talented UX designers only being able to operate within that space.

It may be state of the art, but it doesn't mean that there aren't alternative perspectives which would be equally as successful if given a shot.

My understanding is that part of the interview loop (for software engineers without significant UX experience) is to ask what can be improved about the UX. The correct answer (in the loop, from the perspective of the interviewers, according to my understanding) is “nothing; it’s as good as it could possibly be”. This doesn’t exactly agree with what you’re saying because there are potentially better alternatives outside of Jeff’s ideal (I’m not too worried about it as both of our positions are reasonable).

Amazon is one of the few UI's I don't have major gripes about, so good on him.

Personally, designers SHOULD be more restricted, the problem space itself should restrict what they are able to do, but it often times isn't.

At one point someone decided the shuffle button for the iphone needed to be hidden for whatever reason. I wish the designers had been restricted by someone's personal wishes during those discussions.

> (You wouldn’t believe the thought/effort/testing that can go into the shape, color, details, etc. of a web app button.)

I beg to differ. A/B testing minutiae like the exact shade of color is not UX research; it's merely the pretense thereof.

All the people doing this research are not exceptionally brilliant: this is easy to deduce by noticing how everything is getting objectively worse.

> UX design (I’m not speaking of other aspects of design) isn’t really a creative process as much as an engineering discipline.

Absolutely wrong. It's creative in the sense that you can't quantify it with scientific measures.

UX is coming up with the iPhone when all the smart phones in your era had a small tiny screen and a 4x4 keyboard.

Not sure if this is trolling. Please don’t be mean to me. Your comment is abrasive.

I can assure you that engineering departments do in fact teach classes in which the lecture material is based on research regarding the color and other aspects of design elements. The parent discipline is “human computer interaction”, and “user experience” is a narrowing.

I’ve sat through an hour long presentation about a single button used to meet an accessibility mandate in California. The presenter spent months employed with the purpose of designing the button and designing tests to ensure the correctness of the button. The size, shape, color, and sub-images of the button all contribute to its effectiveness.

Coming up with the iPhone as a creative example of UX doesn’t make me wrong. It’s at the intersection of design and engineering in the same way UX scientists have to brainstorm research directions. My comment was specifically regarding the formal field of UX as a subfield of HCI, and it’s correct and relevant to the subject matter of the article in that websites and apps look the same because there is a state of the art in the engineering discipline of UX.

I used to work in UI (within game development) and there are very good reasons for this. Almost every time someone decided to do something different, usually thinking themselves innovative, it meant disregarding conventions and confusing and frustrating players. There were some exceptions, but experienced designers know when to stick to conventions and the short of it is "most of the time".

The most frustrating experience for our team was regarding the design of a widget meant to convey a number of properties and statistics about some in-game object. Our very experienced Senior UI Designer came up with something which was simple, clean, and looked great.

A manager without UI knowledge didn't like it and mandated that it was replaced by his own unique vision, despite protests. This resulted in half a year of iteration and tests, bumping against problem after problem which needed to be solved. Readability, contrast under different circumstances, data edge cases, you name it.

Meanwhile, our senior designer was forced to follow along, protesting and reminding everyone at each step that he had a perfectly fine solution to begin with. In the end, after months of painful iteration the manager had a brilliant idea, suggesting a fix for the final remaining issue. He had now, by ordering our Senior Designer how to perform each iteration and then evaluating it together, designed the perfect widget. Our senior opened up his original rejected design and showed that it, indeed, followed that specification to a T.

One wonders how much the original could have been improved by starting with that, and then testing. Of course a senior designer should get close to ideal, but taking a good start and then testing normally reveals minor flaws.

Yes, that poor guy was so endlessly frustrated both by the waste of time and the complete lack of appreciation for his skills and expertise.

A more senior designer would have known to convince the manager at the outset that the design was actually his idea.

Usually the designer in question could do that, but... Let us just say that the manager in question was the reason several people quit and at least one person was fired for insubordination. He was not reasonable or easy to work with.

> I used to work in UI (within game development) and there are very good reasons for this. Almost every time someone decided to do something different, usually thinking themselves innovative, it meant disregarding conventions and confusing and frustrating players. There were some exceptions, but experienced designers know when to stick to conventions and the short of it is "most of the time".

Well here's why you're wrong!

hint: you're not wrong. Even as a technical person it drives me nuts when interfaces change for the sake of it.

I've come to dislike designers over the years. Not because I dislike designers, but because it seems like form has taken too much precedence over function and often that form changes as the old guard is replaced with the new.

This is very easy to disprove:

UI design follows trends.

The current trend is flat UI, so everyone is doing it. Before that, there used to be a different trend, and everyone was following it.

These trends have nothing to do with helping users.

Doesn't flat UI confuse and frustrate users? My contention is that it does, yet everyone is doing it.

Constant popups to remind people to sign up for news letters? Frustrating.

Too much negative space and not much useful functionality on the UI? Confusing and frustrating.

Hiding features that people want to use all the time behind a hamburger menu?

Removing functionality all together because the designer couldn't figure out a way to fit it into his design?

Almost everytime some application does a "redesign" to "modernize" their UI, they do it by following current trends, and the result is an objectively worse UI:

- Users are confused and frustrated.

- Previously accessible functionality hidden behind more menus or removed altogether.

- Slower/Jankier overall performance.

Really nothing about it has anything to do with caring about the actual user experience.

Playing devil's advocate here : it is sometimes desirable to reuse a design that your end user's are going to be familiar with.

So there's a tension between "being boring and usefully" and "being original at the cost of forcing your users to adapt".

Case in point : I had to explain a feature yesterday to a user, and it's complex and not trivial, and what basically saved my day was saying : "it works like in X".

Also acts as a self reinforcing thing.

Like games, wasd is movement, so every other game copied that. If you try to change, people are going to be mighty confused why the controls aren't "normal".

WASD movement is an innovation. Movement was on the arrow keys before that.

A wave of games that wanted you to play with the mouse and keyboard simultaneously made using the arrow keys impossible, unless you wanted to use the mouse with your left hand. If the auxiliary part of the keyboard had been to the left of the main part instead of the right, movement would most likely never have come off the arrow keys.

PC shmups, mostly of PC-98 and MSX traditions, never stopped using arrow keys. New ones released this year still use arrow keys for movement and Shift+ZXC for action buttons. Probably because the mouse was never relevant for that genre.

Everything is an innovation the first time it comes around.

Not if your model of the world is "whatever people are currently doing, everything new will do it the same way".

I think WASD came from two player keyboard only games before the mouse existed. Player one was WASD (or something else), player two was the arrow keys, or something like IJKL (it has been a few decades since I played a game like this so I don't remember)

> Like games, wasd is movement, so every other game copied that. If you try to change, people are going to be mighty confused why the controls aren't "normal".

Sometimes though it is for the better. A lot of games in the 80s, especially on ZX Spectrum, etc, used QAOP for movement (QA vertical, OP horizontal) for some reason.

Of course the best is to have these configurable - both the Speccy games of yesteryear and modern games allow you to change the controls to your liking, so people who for some reason still prefer QAOP (is anyone still out there, i wonder) can still use that scheme in modern games.

A sibling to your comment suggests why QAOP doesn't work any more: many games require a keyboard and a mouse.

QAOP requires two hands.

It's also a good example of how copying a design without fully unerstanding it is a bad thing.

A common complaint from the French is that games have messed up movement. Why? The French keyboard looks like this:

    A Z E R T
     Q S D F
      W X C

Obviously, you can't use WSAD on that. As a designer, you have to understand that you're after the placement, not the labels.

Even if you rely on scan codes, you're only guessing the button placement. They can be arbitrary, as long as user's key map matches.

Now you're down to arrows. They have an embedded meaning: directions! You can rely on them being placed in a way that makes sense to the user, even if it's not necessarily the standard WASD positions.

Scan codes should be the main solution, not arrows. Keyboards that change the actual key placement so badly they can't be used are very very rare. Needing several buttons in addition to the movement keys is very very common.

Even if you're worried about weird layouts, the biggest issue you'd see by far is keyboards that don't have arrow keys!

Scan codes might be the main solution, but why are arrows not a solution? If the keyboard has no way to input arrows, I'm sure it's by user's choice, who's very aware of that :)

Even controllers assign buttons a meaning, rather than location.

I already said why. Not enough adjacent keys.

> Even controllers assign buttons a meaning, rather than location.

Not really? There are arbitrary labels that change from controller to controller and the game displays whatever it thinks you have. Like keyboards.

Yes, trying to guess the appropriate version of "wasd" based on either the locale or the actual keyboard layout is a design decision.

But, again, if you're going to say "you can use the left part of the keyboard for directions", then you have different options, it's just probably safer to use wasd for QWERTY, qsdz for AZERTY, etc... rather than anything else.

It's the running joke - every new startup is "Uber for X".

The issue I have with this is that design is typically domain specific. What works for an everyman-search engine might not be a perfect fit for an accounting software, an professional team managment tool or an powerplant control software.

A design targeted to the occasional user can spectacularily fail for power users that use the thing 4 hours every day. Sure it is easy to understand and self explainatory. But this is an advantage once at the beginning and might cost your users every day after that.

Having a steep learning curve is not a feature, but sometimes easy to use means a UX sacrifices fast and efficent workflows (cue notepad.exe vs vim).

Sure, notepad.exe is easier to use (and probably has more users in absolute numbers) than vim, but if your target group is power users, guess which one might feel more worth investing into for them?

VIM should be compared with an editors that has a bit more power. Kate (from KDE) perhaps (there are other choices, Kate is one I've used)? If all you need is the features of notepad, then notepad is strictly better than vim - highlight-copy-paste and save being the only features, and notepad using standard key shortcuts is easier to learn than the vim keys. (note that gvim has menu options for those so you don't have to use the vim keys - I'm not sure if it supports the standard control key shortcuts).

However if you are programming all day vim has a lot of powerful features that notepad cannot touch, so if notepad is your other choice of course it is worth learning. However if your other choice is something more powerful - it turns out there are programmer editors that are just as powerful and a lot easier to learn. The question then becomes is it worth spending hours learning all the commands - including for things you rarely use - for the few ms it saves you each time you use them. (make sure you check with a stopwatch - people often feel that something is faster when the stopwatch says otherwise - as any UX engineer should know, but most vim advocates do not admit)

Good comparison there.

My point was only that some things need to be simple and self explainatory, even if it creates friction and reduces the power, because:

A) they are not used all that often so the friction does not matter

B) the thing is public and must be usable by potentially everyone

On the other hand some things that are meant to be used for hours at a time should make the same tradeoff between reducing the one-time mental burden vs reducing the friction in daily use by power users.

A cash register that is operated 8 hours a day by one person for 5 days a week for years at a time can (and I argue should!) be designed differently than a check out register that is used by new unexperienced customers for example.

If an action the cashier does 200 times a day takes five seconds longer it makes a difference of 17 minutes a day or 83 minutes a week. That is a workday wasted on nothing in a month.

The same workday could have been spent to train the cashier on a faster cash register that has a tiny bit more powerful user interface.

Five seconds are a tad bit much, but even if you take one second, that number would add up to day in five months of work.

And that is the capitalist efficiency perspective — it also annoys cashiers if they are slow to do the one thing they have to do 200 times a day.

And this is a good thing. Apps don't need to be unique or special, they just need to work and do their jobs effectively.

Overstyling just introduces cognitive load and makes it hard to figure how to get things done, especially when designers feel the need to reinvent basic navigation flows and deviate from the boring established practices.

IMO as an industry we should fight back on the user's behalf against needless styling in the name of branding or design or whatever. Nobody cares. Nobody.

> "Why are you doing it like this? No one else does it this way".

I think that's a really good question. Doesn't mean that it won't have a good answer, but if the answer is non-obvious then why shouldn't it be asked?

True, but all too often the reason nobody else is has a great answer and you will spend a lot of time figuring it out. That is time you could spend instead on solving a real problem. Sometimes it is worth your time to ask why nobody else isn't because the answer turns out to give you a competitive advantage, but most of the time they have good reason and you will spend a lot of effort finding that reason.

Well company's usually have brand books which lean into this inoffensive sameness to create a uniform personality. We can't expect every employee to be a creative genius, so the result is that everyone is muted.

Reply on news.ycombinator.com