Why isn't People-Centric UI Design taking off? | Hacker News Reader

Why isn't People-Centric UI Design taking off?

hanselman.com

93 pointsCodeCube75 comments

Communication tools that people use heavily enough to care about acquire a context beyond just "Who I can communicate with". As the recipient, for the majority of my routine communications I get information from medium > sender > content, even though all ways of reaching me just make my phone jingle.

The series of decisions that results in me calling someone is so radically different than the one that results in IMing them that I'd never find it useful to start with a person and then decide to IM or call.

I'm about as often in the task-oriented position of thinking "I should call someone" or "I wonder who's awake on jabber" as the person-oriented position anyway.

tl;dr: The medium is the message

I see no reason that the OS can't support both flows.

Which, incidentally, Android does.

Well,

A phone could have a list of people and clicking could activate the preferred communication method. It might have smart algorithms that even choose the people you'd be most likely to communicate with in the context you're in. So the aim of this UI idea, saving keystrokes, gaining flexibility could be gotten even in your context.

The problem is that even if the aim is achieved, if the keystrokes could be saved and the flexibility increased, it would not benefit someone that much. Most people aren't speed typing, most people already can remember all the people they want to communicate with already, etc.

UIs have stayed at the point they're at based on the "good enough" principle. UIs may not be optimal altogether but optimality only matters when you're on the "critical path" and with a UI, thinking may be the critical path and keystrokes, associations and etc may come a distant second.

Well, the medium is of course the message — but that doesn't mean that we can't encourage users to think of the medium of electronic communication as a single thing. We know it's all just some structured data being bounced around, why not make that clear?

I think that would be a great thing, and would be the main benefit of pursuing the approach the author is arguing for.

Well said.

The thing about computers is that yeah, they're awesome at communication, but they do a lot of other stuff too that has nothing to do with person to person transmission of information. IMHO task-oriented UI makes the most sense because I use my computer as a general-purpose tool to get stuff done. In all use cases I find myself performing a type of task. Only in a subset of cases do I find myself communicating with a person.

I guess I'd call myself more of a power user, but when I look at what I've got going on my computer at night it typically looks like this:

- playing a video game

- talking to arbitrary group of friends on skype over headset

- streaming video to twitch (to whomever is watching)

- programming my side project on my laptop during downtime (waiting for people to connect in-game, etc.)

- browsing the web (looking up game items, watching funny videos, researching some kind of programming thing I'm figuring out)

- monitoring my inbox

- monitoring my twitter feed

Talking to my friends on Skype is definitely direct person to person communication. But even then, I think more along the lines of "1. I want to have a group conversation with Friend A, B, and X, 2. I'd like to communicate with them through the medium of Skype, 3. open Skype 4. add the specific friends I want to talk to to my conversation, 5. go ham". There's no one friend in particular I would "open" first, and subsequently open other friends into. That seems backwards. It's more about what I want to do, not necessarily who I want to talk to.

Exactly. I think the obvious answer to the author's question is that apps are verbs and people (and documents) are nouns, and when people are trying to use a tool it's the verbs that start the process.

The real problem, which the author does not discuss even though it's implicit in some of his examples, is that many of the apps on a typical phone (or desktop computer, for that matter) aren't the right verbs; they're verbs that were easy for the programmer to program, not verbs that are easy for the user to use. You fix that by making better apps, not by switching from app-centric to people-centric design.

In this case, the document would be a shared "group". Skype would be the backend application that handles any interaction within the group, but it's the shared conversation and simultaneous voice/video chat within that group that the users care about. You'd have to go to the effort of adding your friends' "people" objects to the group at first, of course.

The article brings up a good point, sometimes it makes more sense to start with the person. For example, let's say I need to get in touch with my friend to quickly inform them of a change to our plans.

Open Skype? Not online. Check Facebook, nope, they're not on. Messenger? Nope. Ok, I'll open my contact book, and get their number to call.

Wouldn't it make more sense to simply select the person I'm trying to reach, and see what platforms they're online, so I can choose my contact method?

Or what if I want to see what's new with one friend? I should be able to click their name, and see their facebook changes, new tweets, status updates, etc.

Of course, this isn't always the case, I might want to open Facebook without plans to browse a particular person, or I might open Skype just to see who's online that I can video call.

So, there are times when I need to take both routes. They should both be available so I can choose either. It's complicated to aggregate content from a variety of sources into a single page, so it's not a surprise why this option isn't frequently available. You would almost need a new standard, so HN could provide an API for a user's comments, posts, etc, and Facebook, Twitter, Reddit, etc, all do the same. Then I can build an app that merges all of this information together. So a user opens the app, and creates a page for John, enters John's Facebook profile, Twitter/Reddit/HN handles, and everything appears in a single place. However, these sources then get less direct traffic, and therefore less revenue, so it's unlikely they would jump on-board.

From the platform provider's point of view, it doesn't make business sense to offer features like this to an OS or to other apps. Someone like Facebook wants you to have to come to the experience (app, website) that they control, and that they monetize. Want to know if someone is online? Come to our app and find out. And maybe while you're here you'll browse your feed and see an ad or 2 (probably more like 10).

Yes, that's why I said they'd get less traffic, and not likely provide the information. However, you might be able to achieve the opposite.

Let's say on the iPhone, you could click a person's name, and attach their Twitter and HN handle. Now, you click their name a few dozen times a day when you go about your usual stalking to see all the new tweets and HN posts. You then follow these links to both Twitter and HN when something catches your eye. Facebook is now missing out. They didn't want to take part, because they wanted you to visit their app directly, but now, people are not taking the time to open their app. I can get 10 sources of information on my friend by clicking their name. Will I make a completely separate visit to Facebook to get their info? It's kind of a hassle, I might check those 10 sources with a simple click every hour, but only visit Facebook once a day because it's now more work. If Facebook was integrated, and I now had 11 sources of information just by clicking someone's name, I might start visiting Facebook more often by following these links.

Exactly. Global person-centric presence.

click a picture of your friend and then contact them in any possible way using any enlisted app from there

Absolutely not! Not to scream and shout, but absolutely NOT!

I have contacts who are both friends and coworkers or business partners or customers. I do NOT want to get in touch with them with "any app at all", I want the app that supports the workflow or information sharing pertinent to the context at hand. For a friend context, that might mean facebook. For a partner context, that might mean email - or our CRM - or something else entirely - but NEVER facebook.

The context comes first, then the people.

People-centric UI might work, but only once machines are far more sensitive to context.

I don't think that he meant that the computer would use any way possible to contact the person, but rather that you'd be presented with the option to contact the person with any known/configured method.

I'm not saying that I like that method, but I don't think that it would be an automatic thing.

Exactly. I show this flow in the Screenshots. I see a merged view of one individual and it says "text fred" or "skype fred" or "tweet fred." Clean. Any app can enlist with their verb.

Is our world particularly people-centric? It may be just what we are used to. If I want to mail something to someone, I go to a post office or a post box and mail it. If I want to have dinner with the family, I go to the dining room or the kitchen with food. If I want to talk face to face with someone, I physically walk to the location that "contains" them (office/room/house) and talk to them.

People are not instantly available objects that we interact with in the real world, so I imagine the "block" is somewhere in there.

Say I want to browse the web. Or check my inbox. Or read Hacker News. Or get a file off of my Dropbox. Or read an ebook.

None of those things are "people-centric", e.g. they aren't tied to a single (other) person. People-centric UI doesn't make sense because I use my computer for interacting with many users at once simultaneously through websites, mail, etc. If I'm not doing that, I may be interacting with my personal data. And probably least commonly, I am interacting with a single person. So "person-centric" (not people-centric", despite what the title claims) design isn't taking off because I'm not interacting with one single person for the majority of my time when I use technology.

Add yourself as a people in the people centered UI.

We've already had "My Documents" "My music" "My videos" "My pr0n" and similar such default directories in new OS installs for some time. Just kidding about the fourth one?

I would imagine you'd have under Person "25cf" an option for dropbox, ebook, etc. In fact for shared stuff this could be kind of cool, click on my wife and here's our shared dropbox folder (well actually we're more a google drive family, but whatever, I assume dropbox supports semi-private sharing of folders between users, etc). Or click on my wife and then her ebook collection to toss her a copy of a free (libre) ebook.

Given that it is first and foremost my device, or at least a UI oriented at me, I shouldn't have to go inside my self to find my stuff.

People centric design is great if all you do is communicate with a set group of people, however if you're more asocial, then it's just sort of sad and hard to use.

I justify my sorta-bright idea by being an old married guy with two dressers in the bedroom, mine and hers. Do any married couples randomly intermix clothing, like all shirts go here, mine and yours, even the non-heteronormative couples? So it feels "natural" much as there's his and hers dressers to store clothes (and each kid, too) then the phone would have his and hers phone icons to store clothes, err, to store contact info for apps as per the article.

VLM does accept that this could be highly annoying to some people because VLM knows that some people consider it unholy annoying when other people refer to themselves in third person so VLM feels their pain and accepts that as a perfectly valid style disagreement.

> Do any married couples randomly intermix clothing, like all shirts go here, mine and yours, even the non-heteronormative couples?

Yes, and not even sorted by type. This is due more to limited space and getting bored half way through sorting things out and just shoving it all in rather than some thought through philosophy.

The main difference is that you share an actual, physical space with your wife and kids. So it makes sense that you have partitioned places for your stuff. I doubt you share actual, physical devices with your wife and kids, so it makes less sense. If it's your device, the interface can be specialized to you, and not have to present a generalized Your Family interface like your house does.

(You may see a deleted comment of mine; I re-read your top comment and the analogy clicked.)

> Do any married couples randomly intermix clothing, like all shirts go here, mine and yours, even the non-heteronormative couples?

Yes.

Just kidding about the fourth one?

Perhaps the only genuinely useful thing about the cloud is that it means we no longer have to store porn locally.

I assume the "My" stuff is a useless relic from the pre-NT days. "My music"? Of course it is mine, who's else's could it possibly be?

Well... the lawyers want you to think it's not really yours, it's the copyright holder's...

Ditto for "My Computer", "My Documents", etc. I think Vista was the first to get rid of the "My" prefix - and whatever reason they give for it, I still think they want to slowly separate users from the idea that they own (and thus are in control) of their machines...

On my phone I have about 20 apps that I use frequently, 200 people that I need to contact, and about 2000 documents in the form of media files and notes.

It seems obvious to me that application-centric design is the simplest way to organize my computing device. I think pinning people or documents to the main menu would make things way too cluttered and confusing.

But if those people are your spouse and best friend? The apps on your main menu are the everyday workhorses, not the niche utilities. It would be the same with documents and people

I think the article author has a great point, but it isn't a paradigm to overturn all others. Just a position that people-centric should be a first-class use case, up there with people-centric and app-centric.

Let's put it another way - do you frequently open Acrobat or Excel then use File -> Open to navigate to a file? I am absolutely positive that most people click on the file most of the time. In the same way as OSes have favoured apps - phone main screens, the OSX dock, desktop icons, pin to taskbar - there is usually a selection of often-used, easily accessible documents - desktop icons, document stacks, recent files, favourites etc. Why not have your favourite people in the most-accessible areas?

Of course, one answer is that we are still most of all still file-centric, and there isn't a common container for person-with-all-their-contact-details. (And not just their contact details: the person container could have all kinds of indicators of their status and recent activity, at whatever level they have shared with you). But phones are busily trying to squash the file metaphor as it is, and the 'main screen tile' is a general kind of container that can be, already has been, extended to have person-centred features.

I think we will see more of this, and we won't all like all of it, but as a whole it will be a step forward.

I think the issue has more to do with siloed communication than with how people want to interact with each other. Take the Google Hangouts app for example. It uses a similar approach to what the post advocates: click on a person, then select text or gchat as a medium. But I don't really care if I'm using text or gchat, I care that the recipient gets the message that I send without a duplicate message arriving in another medium. Which is how Apple messages works within their closed garden, but it's not going to find my friends on email or Facebook message or gchat or whatever other medium.

So we go to a dedicated texting app or dedicated email app because we assume the recipient is going to get the message from that medium in a convenient and timely fashion. Whatsapp, Snapchat, FB Messenger, Hangouts, etc etc all contribute to the app-centric model by creating new silos. See http://xkcd.com/927/.

"Even on my iPhone or iPad I can't have an icon that represents a document."

You can do that easily on Android, using a file manager, and on a desktop you can do it too. There's nothing difficult about this, it's just not on iOS. It is an interesting point though, I wonder if a people widget for Android would be useful. I seem to remember having a most contacted or similar by HTC a while ago, but there's nothing for stock Android atm. There are of course some on the store, but it just doesn't seem particularly useful. I guess it just doesn't seem easier to me, the task context seems useful.

Edit: Contacts widget = widget for People :)

> I wonder if a people widget for Android would be useful

Is the contact widget what you are looking for?

Yep! Thanks, I hate the Android widget menu so I just guessed it'd be called People since the app is now.

I just put shortcuts to contacts on the homescreen and it works fine.

Fundamentally, it's because we're tool-using primates. For most of human history we've created things to use on other things to do things.

Selecting an app isn't any different than selecting the right tool for the job at hand. In this analogy, your phone' s homescreen is your toolbox - here are all the things you can do. Personal preference and the specifics of the task count for a lot which is why there's many types of hammers and apps with similar functionality.

From a toolbox perspective, performing an action on a contact - e.g. calling or texting - only involves a few out of the many tools available and therefore shouldn't be the sole way to navigate the system.

However some contexts a person-centric UI is great, and one of iOS' failings. I should be able to go to a person in the contacts app and see all the available ways I have of contacting them depending on what apps are installed.

Maybe when AI becomes so good it knows what I want to do before I do, or when we've evolved to the point we no longer feel the need to hit things with other things, a tool approach won't be necessary. But I think it will until then.

Because the UI community has not yet found the perfect universal way to represent people, and it probably never will.

I've often imagined how great it would be if operating systems were a little more opinionated about how data is represented, and how data and tools interact. Wouldn't it be great if every photo on your computer had the same rich capabilities? Wouldn't it be nice to be able to text your friend right from the browser when you see them mentioned in a news article?

But over the years I've seen how UI innovation happens, and the way this plays out is always "constrictive OS UI principles from several years ago" vs "new hot app that does one thing awesomely in a way that's incompatible with the UI toolkit".

And in the end, it's more valuable to people to have a really nice whatever-it-is than to have consistency and interoperability across all of their apps.

For that reason I think data will probably always be represented by a proper name first (Instagram, etc) and then the actual thing second.

Faces?

This is an interesting article. And you will not find many like that one around. For starters, why I think Apple approach is also People Centric. Because software that solves a specific problem, deserves an icon. One solution - one set of documents. You can drag images here and there with Apple API form one app to another if you like to do something specific with them and they are kept under that specific icon. Camera+ is a good example. iAWriter is a good example for texts. You have your documents _for writing_ in the app that does exactly the thing - writing. And I like that. And no one asks why I cannot Open With another program. No orphaned documents in Apple model.

One thing is that different communication programs are different "spaces" - you tend to associate one with one kind of thing for certain people, another with another. People-centric is really just one instance of document-centric, the "people" just being the data associated with the person. I suppose these icons on windows phone with different apps that can be opened from them are basically Windows's "open with". Android's dialog you get when opening a data url is the same letting you choose what app to open it with, such as if multiple apps are associated with phone numbers and you open a phone # from your contacts app.

Because it feels creepy. I don't want to "pin" a person to my home screen because it just seems too attached.

I did have a windows phone and very much liked having my wife pinned to the home screen but she was the only one who made it. I wasn't interested in having a whole heap of other (less attractive) heads burning my retina all the time.

I am probably unusual but I find the whole social media/Skype scene quite tedious and I have closed all those accounts as time goes by. Excluding my wife, replies and others contacting me, I send about 2 personal emails a week, about 5 SMS a week, make about 10 personal phone calls a week and that's it. It simply doesnt justify a people centric UI.

Thinking about it, this is a weak variation of a much earlier failed paradigm. The original idea, document-centered UI, would not have apps at all or rather would have one huge app that was somehow powerful enough for all the users needs, light-weight enough to load quickly and flexible enough to let you link any document to any other document.

If you have that, of course you'd want to start with documents and people.

The problem is you won't find this single app, Google Wave was a failure, etc.

So if you have apps, well of course you do want to decide which to open first.

I met two "new" sysadmins the hard way today (fixing a problem).

People centered would work with people you already know. But the mode switching impedance jump would be icky when you inevitably need to go into search mode. So I'm totally used to clicking on my people centered UI boss contact icon, but now I need some other mode to look these guys up in company email.

Even worse I'm not sure how to represent social distance and rapid changes in social distance in a people centered UI. Yes I've never talked to these friends of friends of friends before today, but this morning, oh boy were they ever the most important people in my social network.

Finally talk about degrading non-gracefully. "John Smith is broken" What does that mean? Then you've got to do the whole "ignore the man behind the curtain" to break thru the abstraction to figure out twitter is down again. Its going to feel very fake when it works and helpless when it breaks. Ah I know just make sure no service ever fails (LOL).

PS am I the only person who looked at the "modern" UI and winced? That is an utter abomination. Hurts my eyes hard to read. Side by side I would never select the "new" UI over the "old" UI. More cluttered than a hoarders basement. More garish and clashing colors than a 1995 webpage or a 1985 color desktop publishing memo. At least the ancient UI could be decoded and read as a text menu with visually annoying cruft, the new one is unusable.

I think that part of it is that the thought process is closer to "I want to send an email to my friend Bob" than "To Bob I want to send an email".

I realize that sounds contrived, but I think activity/tool/app centric experiences are more closely tied to "solving a problem" and people centric experiences are more obscure/less obvious in what you can do.

Perhaps the problem is that people think more about what they want to do than who they want to do it with.

This is also a representation of where the power and profit sits within the industry. If apps were subservient to people in the UI, it would be trivial to replace them. However, successful businesses are built through preference over competitors. These businesses pay the wages of people who design systems and services that achieve this. The app stores continue to feed the platform makers and that's where we end up.

I don't have any friends or know any people. This wouldn't help me much.

You could still list people you stalk, or give machines anthropomorphic names, the ever so stereotypical pr0n star name or star trek character name for each server, etc.

"In a shocking turn of events, it looks like Jenna is, erm, down again...?"

without getting into too much about how I can put icons to documents and people right on my Windows desktop and Android start screens, my amateur psychologist thinks that the reason people-centric UI design hasn't taken off is that we subconsciously associate functions with people. In other words, our interface with another person centers around the functions that person supplies us with. For example, I have a few friends that I only ever play online games with, so my functional connection to them is via games or game-centered social networks. I think "play games with..." first before I decide on a person to play with.

This is why it's natural to chart out social relationships as entities/nodes for people and relations as modes between the nodes. Navigating to another person involves traversing this graph over a selected relation.

I think this is the reason I go to facebook before I think about posting something on a friend's wall, or open up my email client, or picking up a phone first, before finalizing my internal decision on who to call.

I think only because it's harder for the host environment to actually succesfully pull off in a consistent and reliable way.

Which is, I think, the same reason that iOS went so far back from 'document centric' to 'app centric' at first -- because it was easier for them to pull off and make work consistently, both in terms of technology and UI.

> The "story" around People Centric is that you don't think "go to twitter and tweet my friend" or "go to Skype and call my friend,"

Yes. Yes, I do. I use Twitter or Tweetcaster for tweeting. I use Google hangouts for video calls. I use Chrome to view websites. Implementation details not only matter, they matter a great deal.

The market leaders cannot/will not implement large breaking features taken from market losers. It's the same reason why most of WebOS's innovations haven't been adopted, and those that have have done so in a much more muted way.

All of the Android phones I've had in the last few years already do this. Unless I'm missing something? I usually set it up with my favorite contacts on the home screen even I first get a phone and then don't use it much.

The issue is that texting, with a few interesting exceptions, completely dominates most people's daily personal communication. I mean, the "People" app on my Android phone can do what he describes, but 99.99% of the time, my most recent text conversations are all I want to see when I pick up my phone to communicate. Both Hangouts and iOS's messages have convenient links to their phone app, making this pretty much a solved problem.

The only notable exceptions are the privacy-oriented messengers like Snapchat... but I imagine having the messages be silo'd is more of a feature than a bug for a lot of the users.

I don't see this "people-centric UI" idea as anything more than a glorified address book. And I never use address books, so I don't see why I'd be attracted to a people-centric UI.

The way he is talking about interaction is not people-centric, it is person-centric. If you are only ever communicating with one person at a time, then yes, opening up an individual person's contact details is probably the way to go, and for one-on-one text messages or phone calls this is how I communicate.

If I'm dealing with more than one person at a time, which is most of the time, I open up the application, whether it is facebook, twitter, email etc, and it allows me to view and communicate with multiple people in the medium that I choose.

I think the semi-obvious reason mobile OS's are "applicaiton-centric" and not "document-centric" is the monetization of mobile applications. Apps are the big cash cow on these operating systems, whereas "documents" or people or files can't really be monetized in the same way.

The downside of the app-centric thing is an odd sort of discretization of the user-experience via apps. PC's now feel much more free-form in that user interaction is still largely driven by files (i.e. "documents").

"click a picture of your friend and then contact them in any possible way using any enlisted app from there"

First, I think this is a good idea. We already do this with documents on our computer. Sometimes we launch Photoshop first, but other times we right-click on a document first, and choose "Open With...", then pick an application.

But I'd imagine third party developers wouldn't really like the idea of People equivalent of "Open With...", because that takes eyeballs away from their own contact list.

This works for devices that are primarily communication devices, i.e. ~2008-era phones. Today I use my phone to browse the web, check the weather, read the news and tweets, trade some stocks, post a few pictures to facebook, and maybe send an email or text. Oh, and occasionally make a phone call. Of these, only 3 activities are 'people'-centric. The rest are activity-centric and apps are the modern smartphone's encapsulation of an activity.

Collecting the information of the various ways that Alice can talk to Bob is a righteous bitch ATM. However, if we make it easy for people, we also make it easy for machines, and that leads to massive privacy issues.

I like that my profiles are available to people I want to communicate with. I don't like that they are also potentially available to siding salesmen (I live in an apartment).

The pinning feature on WP8 is great, and I wish I could do it on my iOS device. It allows you to create deep-links within apps.

Because those are not real people, they are contacts. Also, computers are tools, they have functions, not people in them.

"In Windows Phone and Windows 8+ I can pin People to the Start Screen. It's a killer feature that no one talks about."

This isn't new, my android gingerbread (2.3) phone does this. FWIW I'm going to upgrade this year . . . maybe.

Yet another attempt at reinventing the wheel, and one that's wondering why people don't like Metro, no less.

Talk about the blind leading the blind. Perhaps everyone hates Metro because it's a bad idea, badly executed?

Because we're still in the stage of figuring out the different and optimal ways of communicating with each other, let alone presenting all choices to a sender.

Because devices are tools and we use tools to do something.

Because it's too German in syntax. I think "call Michele" not "Michele anrufen" or whatever. Verb first.

Well, my activities are not people or document centric. Most of my icons are games, or various other apps.

I could pin people to my desktop on my n900. windows 8 didn't invent this...

That was a great phone. Wasn't saying it was invented by Win8, just that no one is using it.

Because apple hasn't released it yet. rimshot

He's essentially describing Android.

Yeah, I use an Android phone. When I type in my wife's name (or click on her face on my home screen), I can choose to SMS her, Skype her, Whatsapp her, email her, Facebook Messenger, etc.

Because who cares? I think most people just have things they want to do with their device, they learn how to do those things, and that's it. What does being 'people-centric' actually do for me? Does it even save any clicks? How many, one?

It has.

It's called touch screens.

Thats the big UI/UX improvement.

Reply on news.ycombinator.com