Embed code
alright so i think i think we all realising especially
or listening to the things i was talked about yesterday that uh it's action paradigms we've we've
seen before i've really been changing in let's say the last ten years or something
i've if you think back ten years all the things that i've been
in changing um well you were at a a i talked yesterday
it's when you listen to all this your talk about the mass and uh
the technical side but there's a human side to that is i mean we had all sorts of different questions
more about there's a side to side the human side how we can search interface with all this
and it's something we need to figure out how we're there are some few people that obviously in
in companies like cool that are focusing on this but
most the interaction designers most software engineers most
schools they don't really think about how we supposed to interact in this brave new world
but we've had ten years to figure it out and we haven't really started in general
well it is a bit surprising that whenever you on a on a typical i.
t. project we're really just sticking to what we've been doing for decades
it's i'm on the project now and it's yeah it's just for desktop will
do a mobile version later we'll consider that later because right now
we using big screen so well designed for big screens it'll
work really well and ah h. t. monitor and uh
someday maybe journalist will be riding on tablets or interacting with their articles
a mobile phone but that's not that's not the world we live in within the well where you go to the office
and you write the article there and you talk to your call they can have a coffee
and you go back but i'm not really convinced that that's the world we actually live
in i think we just we just used to doing things in a certain way
and we need to start waking up and thinking about how we going to do things differently are we going to make things work
interesting thing is that when it when we first had computers when you first came up
with computers we really started from a very different than point we were asking ourselves
well how should these computers work how should we talk to them well obviously they should try and understand
how human work so they should be compatible with a human speech
yeah it should of china does little graphic that was actually back in the seventies
and if you update that with a bit of colours and a bit of things then it looks a
lot like what we keep on trying to reproduce that was the the xerox out so mark
a computer from yeah whatever early seventies costs
the cheapest one they came up with was like ten thousand dollars or twenty thousand dollars to it
we were all inspired in the beginning by this movie hal nine thousand
was this magic machine in the future which was fifteen years ago
two thousand one that we would be speaking to so with if interface
of the future would be speaking and it would speak back
it would understand or language whatever language we spoke it would understand it
it would understand sentiment woods understand emotions talk about the motions yesterday
well incentives all these complicated human constructs that really drive the way
we communicate the way we think the way we interact
and obviously that didn't go anywhere
because computers at the time couldn't or understand speech any could they couldn't speak
so we're trying to text interfaces instead we asked ourselves well okay
so we can't talk to computer and you can't speak back
but could we for instance ask a question or describe the problem that we have
and then it would it's based on information that it already had been trained with
so just like you go to the doctor unix explain your symptoms and then the doctor based on
all the knowledge that doctor has accumulated knows what process of questions to go through it's
not you don't got just going to the doctors room then you stand there you make a
speech by the doctor that say doctor says right i have the answer here that's
and then you go that's not how it works it's a conversation that goes a back and forth the doctor
asked to clarify and so on so that was basically what they at the time try to build
now we know that it takes a lot more computing power
what the problem that that they had was that compute computers in order
to recognise all the details of the describe problems in order to
in order to match it up with solutions and so on it's not realistic to program every
concert and every question in systematically we have we have to be more clever than that
when we want these gentle solutions like going to the doctor something that that doesn't
mean that you can't do this approach where you have a predictable conversation
and and useful results that you can't have a strict set of that
you can but it's just not for things like going to the doctor or equipment things other things it works quite fine
so today we have a lot of computer car that was um i think early nineties
or something like that that you would have that's a twenty million dollar cray computer
and you'll probably find that your watcher your phone or whatever a device that they use every day we'll
have as much or more computing power than this one did back then sort of of course
there's a huge difference in what we can do to day and so that should
also be a huge different incident to solutions we we take today we
should do things in a different way we should not try to do things the way we would do them when that was a big computer
we should rethink our solutions based on the resources we have available
so let's take an example of typical thing that we probably all interact with every day
but here you has um that's a cool address entry to the right yes
we've all used them where you was a user you have to recognise each of the words
but if it's in a foreign language you don't really know exactly what they mean you have to think about well
if i'm on a particular flaw or to put it in the number or to put it on the street
oh and if i am i have a friend that has a really ought address but she both has uh
later after the number she has of flora end up
an apartment number that's for different potential feels
and which actors entry form will understand that non doesn't exist
that that was entry and and then you into the state first and the
zip code in order to get also complete for the street name
what that whole thing at once did a form like that and
was actually end up spending like two months on is that
on the interactive behaviour also completes and a box and the issues and
so on it doesn't make any sense to do it this way
it looks simple when you sit in the meeting and discuss it but it's not simple to the uses and it's not simple to make
now consider this one over here where you just in principle have one feel to enter in
and you can inter whatever you want so you can ensure a
street name you can industry with the street number first
you can enter zip code you can answer city and as you in
some more information it'll move down when it has understood do
and right out the address so you could start out with the c. d. you could write the whole thing in one line
and you just had some more intelligence you have some more smarts it will take up less space
it will have the same interface no matter what country you're in whatever so you can at that logic
to the context you don't have to redesign you don't have to
have five different designs of what an utterance input is like
or take down here where if you've used um excel or access
who knows these type of applications where you have tables and you can feel to insult them yeah it's
it's a pretty typical paradigm i remember implementing that sort of paradigm in the in the nineties that when it
it was fairly martin that you could just have loads of tables and filter
them and so on but actually it's again it's a problematic interface because
here you have to like these i once did a trading application and we
had an international id and a national id for the same things
there's some obscure numbers and so on and when you see the i. t. u. might not know which one it applies to
so which one are you gonna filter we're gonna put the filter us a user
have to understand ah that number i know that that is this international coed
because it has this format wise that relevant to the user user just has an i. d. for
some stock they want to trade and they should just answer that's talk it should show up
whether that ideas in the name or the idea whatever who cares
the the approach with the search feel that we've learned from the former
decades we google now is should be the way to go
it's the morse it's the smarter into action control takes up less space
it in it's more useful and requires less of the user
complete change of thinking so that was about the you wind the now's about the sort of things that
we this sort of world we're living in as opposed to ten years ago twenty years ago
in the beginning we all say things booked us just a bit of a distraction it's
just it's it's just something to do when you have nothing better to do
and today we have baseball addicts we have people that wake up in the
morning the first thing they do before they got get out of bed
before they do anything ha baseball and before they go to sleep if
after they've done everything and they they just just check face
and it's not even conscious it's just habit not it's not out of and
there's no objective is no i have to do this or have
to do that it's just something that we get used to doing and
we need to start asking ourselves why do we build these habits
instead of renting is that's how that's terrible i should be like that it is like this this is human beings
this is how we are so we function we should start thinking when we design websites when we design ass
when we design products and services or we should ask ourselves why is it like this
what is it about face broke that makes it like that because if you can recreate with whatever you do that effect
you're just fine you'll be just fine and people like whatever it is that you make
whether you think that this is great or not doesn't really matter because if you're
making something that you can be proud of and people like don't deal
then we have chats
cats are actually fascinated because the the other thing that we spend a ton of time with we didn't
used to spend so much time yes we had s. m. s. and that became a huge thing
but traps are way bigger than is a mess because it takes it
a step further has a lot of a lot of advantages
for channel s. is the basic chat where you just chat with somebody
you just having a conversation with another person somebody started conversation by
asking a question hey can you help me out what you
know or have you heard or whatever it is and then you have a little back and forth and then you're done
but the next that is that silicon valley has been spent the last couple of years being really excited about
arts chat parts or conversational you wire or whatever you want to call it which are not new
we've had chat tops i mean developers have known jackpots for quite a long time
and they've apparently also existed with a a wheel chair and so on
the basic premise of a chat about is that it's a counterpart in you can talk to so would
the might have a name or might have a function or something and you can ask questions um
to clarify things or something of that it can also be used to have a conversation those those of
a couple of fate fame is um psychology tests are quite old by now
it lies about was one of the first ones simple little social
experience i'm done by i think it was a stanford's
professor in the linguistics and psychology and it's
it's a program it is not a i it is not algorithms or anything it's
really really early and it's not long it's couple of kilobytes of software
and at the time people were shocked that it seemed intelligent obviously didn't seem intelligent because it's
simplification but chapel scan do interesting things they can also be used for listening in
on conversations and making comments to for instance a useful chat but could be
i checked but that are identifies unusual words it clarifies
the meaning so everybody knows what they're talking about
so that can be extracted to to mean a lot of or to be used for a lot of different terminology or something like that
but i think and more interesting thing is the idea of what i would call chat
apple it's where the message can be into active so consider for instance that you
you in a chat and somebody sends them sends you
a trip that they poke for you and them
where in that little bubble you can actually go and change your seat number
so there you could you could make something like that that means that you you can send
a little piece in as a message somebody else in the chat in the context
we have the discussion you don't have to when you go back to you can find it
by the your memory of that conversation you can go and find that
being can confirm it that way and take the interaction from there
go from there into the application maybe where you can do more things eat it it can
take show license puts it in context it allows you to have a small piece
of information that you can interact with that's quite an interesting concept which of course can be
combined with chat parts so you have a chapel forcing ordering a pizza and then
that it s. a. you just ask i would like to order pizza for tonight blah blah blah and then
it comes up with as it is as it has understood you order and you can sweep the order
it's a very short conversation but it would also allow you to make additional comments
that would be attached to your order and maybe this stuff at the other end will be able to
respond to you directly so that would that's an example of a combination of these
through that i think would probably make for a very constructive user interface
bear in mind that a lot of these jackpots are would be honest terrible and they probably don't they're
rushed out the door the user interface that into action of it is really poorly thought out
so i think and faced doctor now already ten thousand jackpots after they released it
i don't remember what it was six months ago or your goal whatever and that's typical i mean we sort
with with baseball teams and so on things have a certain hype cycle where they boom and bust
but it's not gonna go way but we need to figure out a lot of things
we need to you lot better at making a useful experience with these things
chant engagement is really interesting because it's personnel
if you thinking in contrary how typical way of making websites
then not personal because we're not really having that two way relationship building
you really need to v. p. uh uh put a lot of effort into website in order to make even a personal
feel where you build the relationship that's hard work and it's
something that's really don't you have a two way conversation
and you have the focus of content because it's whatever you're talking about now
think about it when when you have a conversation at any given time you only talk about one cop topic
you never talk about two topics in parallel at the same time but we do it all the time on websites
all the time side bar as distractions ads by different conversations on the same page
and it's not constructive we'll meet to have one conversation at a time with one topic
that's what our brain is good it was good at focusing on one dialogue
and it's a way to get information it doesn't have to be a conversation between two people it can
be extracted in many ways and in here in this example we have an a full or uh
it says something about what it is and the delivery time and that it's
pain and simple doesn't take up a lot can easily fit on any mobile phone doesn't have to be squashed in
you can put it on the website in a nap whatever it would fit in anything
that interface
yeah yeah we have a bit of some other sketches you could have a doctor's appointment there
welcome to alternate you haven't upcoming point you just go and visit
you just go and visit to clinics page yeah whatever or
destination the clinic has and you have a little message that
your next appointment is next monday twelve to thirteen
and it's here on the on the map so you can click the map you can she take the time to reschedule
um and if you would like to know something around your point when you can
just type it and it says we'll read which typically respond as fast
so you have set some expectations there's no distractions that and
nothing to take away from it but you still
it gives you additional control gives you reason to use this over picking
up the phone waiting on a call line and calling them
so that's jumped shoe were navigation
one thing what i looked at the checkpoints that really struck me and
that's probably wanna start it's yeah idea for it for this talk
is when we to navigation on the web or applications we tend to have all these
menu structures we we have a we start out with a very simple menu
um with five or six topics and then as people come in and
want to add more information to the website the menu grows
and all of a sudden becomes a monster and we we say all we need to we need
to high it we need to hide all these extra things because they're not so often lead
use the still need to be there because you never know
with people want to to go from ordering potatoes
well uh straight to to checking out the weather i mean that might be something you really
want to do when your in potatoes to see what the weather's going speeds more
it's we fill up the pages with navigation is completely irrelevant for
what the page does it it happens in most cases
so on mobile we struggle right obviously because mobile phone is not meant to
have so much stuff which we still try to cram it in
the chat interfaces super simple you don't have needed back button
if you want to find something that you looked at before you just scroll up and if you don't find it useful for that
that's it and you roughly remember because we're good at this sort of sequence and spatial
relationship you roughly remember how far you need to go in order to find
and the things and when you see a particular point in the chat you remember that chap and you're like yeah yeah
i need to go up a bit further no that was that was before so i need to go get down
you also have the fact that the most important stuff is at the bottom and the less important stop is for the op
because obviously what you talked about yes that is not as important as what you're talking about now
and what they just said no you can actually have some
also complete you can have suggestions you can have easy
ways of continuing the conversation without sitting down and typing because
typing on touch devices is not the most convenient
typing in general should be avoided if we could just press a button instead
that would be much better it's faster easy you don't have to
uh you don't make any spelling mistake she would it's quite clear what what it is that you're saying
and then of course down at the bottom it allows you to say anything that you want
it was a bottom yeah no sideways scrolling another interesting thing you won't find sideways
scrolling in in in chat you don't really have that conversation there's just
one direction and the direction has a meaning it's not random things that are
connected or you know what you're going to find when you scroll
and no need for back button oh okay so i missed out what one
interesting little thing if you've seen the reason updates to apple maps
uh you'll see that there's the search box has moved down from being on the top
it's sort of in the lower but at the top of the lower third with suggestions below it
which is much easier and it's sort of changes the order of priority where you you actually the
the it's like here and then you have the map out there there's nothing to
press up here in the top which is for our way from your fingers
and then down here is where you would take in order to type and then when you
start typing the keyboard comes up just below it just like it would an agenda
so that so voice
so in the beginning of of this whole whole thing back in the sixties and seventies we concluded we can't
make a hell nine thousand computers don't understand voice they can't speak and so on but that's changing
a voice recognition is actually becoming a usable we as humans we don't like
when we say something in the in the computer just act weird
we don't like that and it only takes a couple of those experiences before we check out and say
no not gonna do that one oh and and it's hard enough as it is um
when i when i go on holiday i could use an audio interface driving the car
but then i might wake up the other passengers that might be sleeping
or they might say well what if talk about why you what's what's going on right
it's it can be very awkward and very we it but
that being said it is a simple useful means of
well it's acting when you don't want to use your hands like you're driving or you
don't want to uh a virtual eyes you don't want to look somewhere else
you want to focus on what you doing which is driving but you
still want to have the ability to make additional choices any sections
or if you just think of something and you want to make in and out so you just want
to dictate something quickly but it has to work harder percent or ninety nine point nine percent
and not surprisingly the or pick tech companies they are each going
for making the voice assistant and so on and so forth
it is definitely something that we'll see more and more of any of that it's definitely something will be more used to using
uh but i think we typically underestimate how hard it is to do think these
things right because working most of the time it's the same it's not working
then we have one other thing that has changed in the last decade it's the size of the screens we use
like i said before we try and design websites like this which is
the they just take it take website they had a redesign
remember your go two years ago so it is a reason design
but if you if you look at how that will render on the phone really probably
so how would you propose to squeeze all of this information in that screen size
it's a really hard job
yeah you have some really frustrating meetings and discussions about how we're going to
remove stuff that you know all some users will want to use
yes but you only use the phone when you are walking on the street no you don't
okay but then you only use it when you outside the house uh no that's not true either
okay most people use of phone they don't use that habit
but yeah but how how many of the users are you going to in it how
many are you going to say yeah you don't might automatically that okay so
there's one percent over there we don't care about there's one percent over there and
before you know it you don't care about a third of all your users
so we need just to start this obsession with boxes and boxes and boxes
and come up with better interfaces that will work across a wide range
of devices and uses scenarios both with voice and with visual the
this website how well will it work with somebody that line
right it's not gonna work very well well the chat interface work was somebody's blind yes
not that complicated to think about how you would make it work as much less
in some information at a time information has a context it has that order
it's much much easier to reasonable when you wanted to work everywhere
i might remind time
so what what the design process be for this that's the big question
i don't necessarily have that have the right answer but i have an ounce
the first thing you need to get right is all
sorts of questions around personality and and pretty
when you are having a conversation personality is crucial
when you're talking to somebody how they come across what words they use what intonation they use
what colours they use the topics that they pick all of that stuff is what you judge it based
on you don't charge it based on whether there's some pretty graphic in that conversation that's not
that's not what you what you it's not the communication you are looking for
you need to think about how you open conversations conversation
openers what's the starting point for the conversation
and how to lead the conversation
because if you have a random conversation with somebody it's very hard to
reason about if you can talk about anything and you're jumping around
if you jump around you're probably going to make them stressed
if they jump around the probably going to make you stress then if you are computer you're going
to have a problem because you're not going to be able to make these human leaps
of of association you need to focus the conversation so the
opening um that conversation opener sets the tone for
what you're talking about and what is acceptable to talk about so in the example of the doctor's appointment
if use just it should use your next appointment is you
sort of say this is to talk about appointments
this is not to ask what's wrong with me doctor could i i'm having a
pain in the left side what would you please tell me about that
so use didn't start out with a blank screen because then you don't know what you need to know because they could be saying anything in return
you want to think about the whole flow this is not just a question of chat
interfaces or something similar this can be seen in the in the peak are are
you think of frame of when you are an organisation you have a personal website or you're making an at
or if you have a service you made a product that has some
digital part or you've had an instruction manual with your physical product
or you have an office with people agree didn't reception you have a support hotline
you write letters to then you write emails it's all part of a conversation that you're having with that person
if you talk to the doctor and you've talked to the receptionist
and you go to the website it's the same conversation that you're having with the doctor's office as a whole
so you you need says you need to have ah
coherent plan for how that conversation should go on
wrong before they become your customer or use or or whatever
when they first introduced to you they have an expectation
a base that on how they were introduced in they'll go through on there
boxes and arrows is an interesting ah i'll raise for
that we'd like to design boxes we draw boxes
and then we say okay so here is the website one page and here's the about page
okay is that transition between their what makes you go to the about page why
you're going there what are you doing next what did you do before
and then you need to think about what the history hand traded
yeah um how you're going to handle their conversation history
showing maybe the history from last time you chatted next time you
come back um how you're going to interpret that history in
in terms of past events and maybe you can have a coherent history
of interacting with the user across all sorts of channels you communicate
the first thing with personality is on it in a a us design
we have this a concept called put the us owners and
the first time i tried out for so long as i was like i'm not sure
i think its own as our heart and easy at the same time
it's easy if you have a really well defined persona then you just use it
then you just discussed okay with this person would work for this person
it's a good persona is somebody you get it when you skim through it and you can visually imagine
what sort of person is you can emotionally imagine what sort of put business you have
really simply ideas of what they what work for them and wouldn't work for them
so that's something to look into if you haven't tried it that's something to discuss but
if you ought to design conversations you need to know who you're talking to
if you don't know who you're talking to it's impossible to define vocabulary
or troll of communication or relevance or any other things you can
come up with you need to know who you're talking to
and equally you know need to know what sort of has a nice of you have
as an organisation how you want to talk to them you want to be formal
informal do you want to be slightly boring you want to be slightly probably
you want to be joking you want to be it has
so many possibilities and you can get it terribly wrong
i'm good example with a check part was um it it
still had a bug because you're suggesting three ounces
and when you click one of them it happens to not understand that on so so what it
sure what it's showed was the sort of for all four of chat parts which was
sorry i don't understand what you just said so i'll show you a picture of a kitten
okay what i get from that as a user is that you don't care about my time
you think i'm here to just play around and i don't really expect anything from you
you're not really planning on providing any value and i should not be here
that's and this is the sort of thing we need to get right
if we don't get this right it doesn't matter if we have all sorts of cool
technology that can do all the ninety five percent other things in an amazing way
we need to really we need to understand what's the person we're talking to and what sort of person now we
and what's the what's the setup for this whole thing
this is what i think most websites look like like if we look at the personality of the websites
they just completes gets a friend x. it uh it it's made by twenty different people
with her into different opinions they all designer for themselves are not really thinking about
what is going to be like using this for somebody else is like well i would like to to i think
and then somebody else comes along is like no i don't like that colour let's change that the deal
is it we we need we need a coherent personality more than we
need fancy colours or two kittens or funny jokes or whatever
we need to get the opening we need to open the conversations in in
good ways um we all know conversations when they get we had
uh oh when they go off track or whatever and it's i just i just think we we need
to think about this i'm not entirely sure about the details but it's something along those lines there
oh it's a good one of them are okay so that's that's it for me
i'm i'm preparing a book that i hope to have ready next summer
and we're going to make a seminar in zurich where we uh take even a wider uh approach and with
are we look at a technologies for communicating directly with customers
the idea that somehow we should now have the tools and technologies are having a direct conversation with
thousand tens of thousands with large groups of people with out
requiring at present every time all the time at
the other end but still having the feeling on both sides that we really communicating as human beings
but any questions

Share this talk: 


Conference program

Keynote
Jean-Baptiste Clion, Coordinator DevFest Switzerland
26 nov. 2016 · 9:40 matin
How to convince organization to adopt a new technology
Daria Mühlethaler, Swisscom / Zürich, Switzerland
26 nov. 2016 · 10:14 matin
Q&A - How to convince organization to adopt a new technology
Daria Mühlethaler, Swisscom / Zürich, Switzerland
26 nov. 2016 · 10:38 matin
Animations for a better user experience
Lorica Claesson, Nordic Usability / Zürich, Switzerland
26 nov. 2016 · 11:01 matin
Q&A - Animations for a better user experience
Lorica Claesson, Nordic Usability / Zürich, Switzerland
26 nov. 2016 · 11:27 matin
Artificial Intelligence at Swisscom
Andreea Hossmann, Swisscom / Bern, Switzerland
26 nov. 2016 · 1:01 après-midi
Q&A - Artificial Intelligence at Swisscom
Andreea Hossmann, Swisscom / Bern, Switzerland
26 nov. 2016 · 1:29 après-midi
An introduction to TensorFlow
Mihaela Rosca, Google / London, England
26 nov. 2016 · 2:01 après-midi
Q&A - An introduction to TensorFlow
Mihaela Rosca, Google
26 nov. 2016 · 2:35 après-midi
Limbic system using Tensorflow
Gema Parreño Piqueras, Tetuan Valley / Madrid, Spain
26 nov. 2016 · 3:31 après-midi
Q&A - Limbic system using Tensorflow
Gema Parreño Piqueras, Tetuan Valley / Madrid, Spain
26 nov. 2016 · 4:04 après-midi
How Docker revolutionized the IT landscape
Vadim Bauer, 8gears AG / Zürich, Switzerland
26 nov. 2016 · 4:32 après-midi
Closing Remarks
Jacques Supcik, Professeur, Filière Télécommunications, Institut iSIS, HEFr
26 nov. 2016 · 5:11 après-midi
Rosie: clean use case framework
Jorge Barroso, Karumi / Madrid, Spain
27 nov. 2016 · 10:05 matin
Q&A - Rosie: clean use case framework
Jorge Barroso, Karumi / Madrid, Spain
27 nov. 2016 · 10:39 matin
The Firebase tier for your app
Matteo Bonifazi, Technogym / Cesena, Italy
27 nov. 2016 · 10:49 matin
Q&A - The Firebase tier for your app
Matteo Bonifazi, Technogym / Cesena, Italy
27 nov. 2016 · 11:32 matin
PERFMATTERS for Android
Hasan Hosgel, ImmobilienScout24 / Berlin, Germany
27 nov. 2016 · 11:45 matin
Q&A - PERFMATTERS for Android
Hasan Hosgel, ImmobilienScout24 / Berlin, Germany
27 nov. 2016 · 12:22 après-midi
Managing your online presence on Google Search
John Mueller, Google / Zürich, Switzerland
27 nov. 2016 · 1:29 après-midi
Q&A - Managing your online presence on Google Search
John Mueller, Google / Zürich, Switzerland
27 nov. 2016 · 2:02 après-midi
Design for Conversation
Henrik Vendelbo, The Digital Gap / Zurich, Switzerland
27 nov. 2016 · 2:30 après-midi
Q&A - Design for Conversation
Henrik Vendelbo, The Digital Gap / Zurich, Switzerland
27 nov. 2016 · 3:09 après-midi
Firebase with Angular 2 - the perfect match
Christoffer Noring, OVO Energy / London, England
27 nov. 2016 · 4:05 après-midi
Q&A - Firebase with Angular 2 - the perfect match
Christoffer Noring, OVO Energy / London, England
27 nov. 2016 · 4:33 après-midi
Wanna more fire? - Let's try polymerfire!
Sofiya Huts, JustAnswer / Lviv, Ukraine
27 nov. 2016 · 5 après-midi
Q&A - Wanna more fire? - Let's try polymerfire!
Sofiya Huts, JustAnswer / Lviv, Ukraine
27 nov. 2016 · 5:38 après-midi
Closing Remarks
Panel
27 nov. 2016 · 5:44 après-midi