 We don't want then your choice, we even bother my experience with... Possibly, possibly. Okay, well we'll get started because we are... It was ten passed. Okay, so something that you might not get to know, get to see much. You're going here on sort of ACI bit research and stuff. That's used to understand it, because you might be asked about it. UX and HCI researchers who are hardcore know about all this, you might have a know of it. Although you probably won't be using it for the most applications, is this thing of user modelingÖ Now, the holdegrain of most scientific application is prediction, simulation of predictions. magnesium mwy o cael ystod yn ynnu'n ddefnyddio i fod yn yn adrodd. Byddai'n rhaid i chi i wneud hynny, bod yn adrodd o sefydlu ar y cyfnod am gwyfledd. Ond mae'n gallu chweithiau chi fod yn byddai'n dda, byddai'n ddefnyddio i fod yn though yn gwneud bod ni wedi cael y ddefnyddio i fod yn arwahanol byddai efallai fel y gwaith yw fy modol? Byddai'n ddefnyddio i fod yn imbiad? byddai'n digsee'n gwneud ffoeddiaethau i'r ffordd? a chynnyddwch chi i gyd, hangi ddod yn fawr a hangi ddod? Yn ddiddordeb am ffondi'r cynhyrch o'r cyntafol o gyllid amllunol? Mae'n sylwod, ond mae'n fawr i'n gyd. Dyma'r lleol yma mae'n gwybod i'r cyllid yn fawr i'r barod o'r oledd i'r cywfwyr yma, mae'n fawr i'r cyllid am fawr i'r barod. Fy oedd yma'r model o'r... Ify, mae gen i gael gyda'r meddwl yn man ρeithasio. Mae'r moddan yn cael maes cyhoeddiol. Mae'r modda i'r cyhoeddiol, lle mae'r modda i'r modda, yn cael maes cyhoeddiol wedi'u gwelwch eto. Mae'r modda i'r modda i'r modda i'r modda i'r modda i'r modd, i'r moddyl i'r modd yn ei wneud... Fyg yw y gallu ddwy'r amser wedi cyflwyno dylunio ar yr wybod. Hwyl mwng o'ch cyflei Adegwyr ffynol i fod yn cymryd, wrth gyda'r newid i'r ddweud o'r adegwyr yn ychydig i gydig o'r ffordd g Paul. Nid nid yw wneud i'r ddweud o'r ddweud, o'ch cyfyrdd yn gyntaf? Nid ddweud o'r ddweud o'r adegwyr? To learn it, et cetera, et cetera. And this is why, this is what was created. Now here, it's generally lots of rules. So for instance, we have things like, okay, so we have a processor, each processor has a cycle time, each memory has a decay time, for instance. So we model the human like they are a computer because we have this idea of a memory decay time, we have this idea of a processor, decay time. byddai'w sy'n bwysig. Mae'r proses sy'n bwrdd angenllunol, mae'r proses sawl o'r proses, a ddaodd yn gwybod. A wedyn byddai'n bwrdd sy'n bwrdd... Yn y bwysig, mae'r gydweithiwch ar y cy harvesting, mae'r gydweithu'r gydweithio, mae'r gydweithio'r gydweithio I mor oedd y blynyddol i'r tylau sydd wedi'i gweithio. Rydym ni'n gwybod plant yma, byddwn ni'n bwysig plesio'r systemhau. Y rhan bryd y phryd a fyddion yw'r ddydd yn ddad oedd, sydd bod ein rhywbeth yw byffeniad, mae'r dyfynodau yn ei фryd i gyda. Mae hynny y gallbwydd yn fwyaf plesio'r molliau, ond mae'r bobl yn ni 5-10 ysgriffad o ddwy. Hefyd, os yw'r dyma, mae'r prysio ei bwysig ei bwysig, goes, goes, operators, methods and selections. Okay. Dyna alla'r ysgol, ac oes bod a'r syniadau wedi'u eisiau iconbolaeth i allanio'r method and selection, might be able to get something more accurate? But again, not by the case. So, remove on to this. The keystroke level modelling... Mae'r Modell Cymru sy'n byw hwn yn fynfer ar gyfer yma, a hynny'n cynnwys. Felly mae'n fynnedig ac yn fyddiw yw'r Tyllid Purwyr a'r Hyd Elu, ond mae'n byw hwn yn fynnedig ac yn fyddiw. Yn cael ei gweld allan, mae'r Rhyw Llyw Llyw taster i mewn. Felly mae'n gobl, mae'n f VMware a nesaf. Fe ymgyrch yn fyddiw, mae'n gobl yn fyddiw o'u hoffod. Mae hyn sy'n gwybod a'r cyfnod o'r Lhyw Llyw yno'n sefydliadol. Er, not very measurable at all. Gones more measurable but more constrained. KLM, very constrained. But measurable, more measurable. Now after this comes this thing, cog tool. Cog tool we've been working on for years. Still not complete but it actually works quite well. You can use it and you can download it and you can use it and people do. Felly, mae'r coctel yw'r CMU, ond mae'n cael ei wneud. Felly, ydych chi'n gweithio gyda'r coctel. Felly, mae'n gweithio'n gweithio'n gweithio'n gweithio'n gweithio'n gweithio'n gweithio'n gweithio'n gweithio. The obvious thing is ask for us there. We actually get to see this keystroke level more. We are convincing, for instance at this frame which is obviously drilled down into the frame that we have just selected that is going to happen. The highlighted step occurs at a certain point. So this highlighted step here occurs at a certain point within this particular part of the story board. Blwyddyn ni'n meddwl yw'r ystyfr hyn yn ceisio. Mae'r nghymru taw'r pwysig hwn yn y d flesh. Dyn ni'n meddwl, rydyn ni'n meddwl ydych chi'n meddwl yr oedd yn symud y stefn y basiwyr eich間u. Fe wnaeth ein ffraeg yn roi'r cyd-farn. Dwi'n dwylo'n erbyn bod yn rhywbeth. Roedd y gall사� o fewn oed? We get once we're done all this, you understand how long things take, how much time they'll take, how much time to take things to occur, and what's going on. W we're waiting for itγArth des מ� sounding like A motor, procedural call, a control call, cognitive calls Anytho w意 siwfio yma iti cei na efo mewn modify yield on the user end, east hee ymwneud unsiwn. ond we've already got this keystroke-level model of the user. What we're trying to do is link this thing into the keystroke-level model of the user with the interface and the interaction scenario. You have to make explicit the interaction scenario you want to see, but we already have a model of the user. Therefore, we can do things like how many things are going to take, what's going to be presented, what should be presented, and how we can make things less there. How could we modify this? It allows us to say, well, if this takes seven seconds, we want to change that now in control, so it only takes four seconds. How do we do that? It allows you to answer these kind of questions, but they're very simplistic questions, and we're just going to point and click on the questions, not overarching questions really. It's still quite good, I think, anyway. This kind of stuff you probably won't be doing before, standard UX stuff, because it's quite in-depth, and you have to do something about it. You have to really, more than it will. You're very well-making for certain aspects. We'll be using this kind of stuff, certainly at Bowie, and the user at NASA, at the American Air Force Academy. This kind of stuff, where you're actually building interfaces for both pieces of kit which are critical. Ken, one of the simplest examples would be how many clicks it takes to get to perform a sort of end-to-end task. That's not what this is trying to show you. This is trying to say how long will it take for a user to get from one end of this interaction to another end of the action, and that it has some cognitive view. So, if it's about clicks, for instance, if it was going about clicks, I could put everything on the page, and it could put everything in one click away, but how long will it take to find that one thing that's cognitively in the touch? Multi-making in multiple clicks. So, when we used to have this idea where we used to measure things in clicks, so everything's only one click or two clicks away, that doesn't make any account for a complex interface which you have to then cognitively navigate. You still have to do something. This new idea of progressive disclosure where we have multiple... We're not so bothered about the number of clicks, but we're just bothered about how logically things fit together. The kind of experiments that we can see that multiple clicks, which are nicely presented with less complex information, allows a faster usability time, a faster interaction time, times test completion time, than things like clicks, number of clicks. Because you have to sit there looking. Okay. So, we can see how this all nicely fits together. Yeah? Okay. So, as I said before, I'm not going to spend too much time on this because I want to get on to the real rules. The real principles are what we're going to learn about. What we're going to learn about. Here, there are plenty of usability principles. Okay? Plenty. So, all I've gone through is looked at the different kinds of usability principles in the nature of IAM page 158. So, I've been through and looked for lots of different usability principles with all the general big experts, the names that you might see, and I've collated them together so you can see where they fit. Okay? And then, there's a rationale for why I'm chucking ones away, or I'm keeping ones, or I'm amoglating ones, or I'm calling it something different. Yeah? Under this. I don't propose to get into that right now because we haven't got the time to be honest, it's not that interesting. Yeah? This is just for you guys, so that you know when you go to a company and they say, oh, I've got a bunch of other members, eight common rules, hey, go to collated principles. Yeah? And I'll look at that and say, oh yeah, I've got a bunch of other members. I don't need to make the other handling simple. That's the same as I say, isn't it? Normal. I've got some normal. Done normal. Yeah? So, this is the kind of stuff we're looking at. We see that they go on, so we've got ones here for safety and short cuts and simplicity, and all this kind of stuff. They're just collated self-principles, which we don't like on these. And these are the sources of those, indeed, of those principles, so that you can go and grab these sources if you have a need to. So if somebody wants to say, oh, tell me about, you know, we're using Benjamin's stuff, well, you just go to the first one in the list. We're using Don Norman's stuff. You go to the third one in the list. OK. Well, this is how these principles fit together. And you see, there's nothing that new. They're all overlapping. Massively overlapping. So, I have my own set of plotted principles, which have evolved from the collated plotted principles. So these are the principles that I think we ought to be thinking about for better efficient user experience. OK. So, some of these are reasonably straight forward, which I'm not going to get into directly because I think they're reasonably long. If you want to do it, I don't view these. Some of them they won't have seen before. They don't really appear in that many overviews of usability. So the first thing, stability. So are the interactions stable? Well, do things fall over? Do things do what you think they should do? Are you able to recover from unstable aspects of the system? Now, I'm going to listen to usability because, yes. OK. So, it's got things that are going to be less usable. The task completion time is going to be worse if the system isn't stable. But really, you could put stability into some kind into the development with robustness, which also is in there. You could put that into principles of development, UX development, say, with agile when you've got stability. Because it's really a part of the whole system that if your system isn't stable, you're not doing a very good job of software engineering for a start. Really, it needs to be reasonably stable. Scalability. So this is a key one. Oftentimes, nothing is placed and it's data is scaled up. So that's a couple of questions. So first of all, it might be that you have, we just did some work where they started off with small sets of data, small data sets. You'll be getting the report we gave them to the system coordinator box. It's part of some of the notes that are part of the appendix so you can see what we actually said. But their data sets, they start off with a small amount, a small set of data, and they didn't want to display it in tables because tables don't look that good. They started to display them in nice JavaScript-looking pop-ups sort of thing. So that's four models, actually. So they look quite cool when there's three or four. But when it scales up to 36,000, then it's just a load of bubbles all down the screen and you can't do much with it and it doesn't sort, et cetera. So the way to make it stable is to put it back into a table. So therefore, you can have flexibility if you wanted to. You can say, up to 10, we have this nice bubbly system which looks cool. But once you get to more than 10 data sets, we put it into a table so you can easily see it. Now, it might be that you only want to see it in a table. But that's the kind of thing. Does the data scale in the interface itself, or doesn't it? The second thing is, does the system scale as well in table? Because obviously that's an interface question because it affects task completion time. And sometimes this won't be the case. For instance, there used to be a shop, an online shop 10 years ago. I had no idea what it was called, but it used to sell clothes. And they had the best servers and they had the best development platform and they made their excellent system and they demonstrated it to their managers and they loved it. And then they put it out and it didn't scale because as soon as it had to do any kind of up and down with a telephone, what would you do that time you did it? Anytime it had to send any data to them so much of it that it was just grinding. It took 40 minutes to go page down. OK. So it looked really pretty when it got there. I'm writing 40 seconds. It took 40 seconds from page down. I'm writing 40 seconds from one page down. So the whole company was based on this internet phenomenon and that was because they couldn't sell enough product because nobody would wait 40 seconds for a download. Why would you? Because the developers haven't checked whether it would scale and because the developers and the user experience people haven't checked how long people will wait for a download to occur. What would the cost benefit of this to them? OK. None of that have been checked. OK. Simplicity. Well, we've been through simplicity with the studies which I think that's reasonable. Situational awareness. So it says up here, is the perception of the interface to facilitate a decision making area? So anybody else want to tell me about situational awareness? Has anybody heard of situational awareness as a term before? Yes. I've heard it in the corporate communication meeting where you have situational awareness that's picking up those pointers whether your message is getting across to them or not. Just picking up those pointers whether they're nodding or not dealing with that. Yeah. That's one area. Any other areas that have heard of situational awareness? No. So it's used all the time in aircraft and flying. It's used all the time. Situational awareness is used massively because you've got lots of difficult tasks occurring more at the same time. You need to know what the situation is, what's happening with the plane, what's happening with the fuel loads, what's happening with the area around you, what's happening with the communication. So you need systems that allow you to understand what the situation is. Now, you might find situational awareness in any of the principles that from Schneider in the library is one that I think it's important based on my research. So you might want to put a line through this in your notes now saying it's heartless, ego-centric witness and you forget about it. But I would suggest that when you're building something, what you're trying to do when it's very important to test is enhance the user's situational awareness. You're trying to allow them to understand where are they in the path. For instance, I would argue that in Amazon, there's a little bar in Amazon which is the shopping cart progress. That allows you to understand something about the situation that you're in, where you are. Are you going to pay? Have you placed the order yet? What's in the basket? This kind of little thing is very simple, but it allows us to understand that situation more easily. But this is very, very important for those of you who will be going on to a large, safety-critical system in more safety-critical systems where users are in the mix. You need them to understand the situation and their situation, be aware of that situation. Self-description. So this is another one which has kind of been hinted at. So where do we do self-description already when you guys already do self-description now? Yes. In documentation for code? In documentation for code. So you guys have generated Java docs, haven't you? Yeah? And so Java doc can only work if you've self-described the code as you go through. What other kind of... I mean, the self-description, what the kind of ways have been proposed for really self-description by other illustrious speakers in computer science or a specific illustrious speaker in computer science. Anybody heard of literate programming? Okay, anybody heard of Nuth? The Art of Computer Programmer? I don't know the other notes. Grief. Okay, so he proposed this system when you write what you take, of course, et cetera. And he proposed this system literate programming when he writes the programs in a literary style. So everything is self-described. The program is its own documentation, if you like. So this is more difficult to do with interfaces without them getting bulky. But what can you do when you use metaphor? You can use understanding from the real world to help implicitly self-describe what's going to happen. Yeah. This is very... It's very tricky, though, with self-description. I mean, the reason why we all resort to documentation is because self-description is very tricky when it comes to interaction design. Yeah. Even the simplest product is a sheet of paper with something like this. Normally. Just to get it going. Let's think about this in much more detail. Okay, a new one, which is not only usability, actually, but it's only really been, if you like, in vogue for the last few years, is this aggressive disclosure. As we were talking about before, everybody writes this. Click time, number of clicks away used to be the way we measure things. The less clicks away, the better. But that doesn't work with complex systems. Okay? Because they're complex. And everything will be on its screen at the same time. And you'll have no situations of awareness. Yeah? If all tasks are happening together, then you also might be able to self-describe it. Yeah? Because they're all presented at the same time at one thing. So, we have this progressive disclosure whereby you can describe a system by what comes after it, by this hierarchy of interaction. What does this say aggressive disclosure aspect and self-description in some ways? What might they negate that we've talked about before? What behaviour might they negate that you might want to see from the system? Emergent behaviour. Emergent behaviour. Because you're already predicting what people want to do when you're putting it into pathways for them. But how do you know that really is the case? And if that is the case with some of the people you've tested, how do you know there's not a bigger behaviour coming? Because you need some level of structure, some level of, yeah, of structure and interaction to understand what the emerging tasks are. Yes? How exactly does self-description and progressive disclosure negate this emerging behaviour? Okay, because like for instance, so progressive disclosure means that you put things in a hierarchy if you like an interaction so you might describe what you're supposed to start with and then you might expand, when you click on that task, you expand out and expand out into more complicated and more specialised tasks as you go along. Which means that if you've placed them into a hierarchy that's reasonably fixed, you can't see the emergence of people doing things in a different way because they might have different hierarchies or pathways to you. Yeah? And we can't test everybody. In the real world, this is writ large for instance, the MIT building that was built in the early 90s, early 1990s was built without any pathways to it. Okay? So it's just a big old building sat there, no main computer block in this big grassy area. So am I going to be into MIT? Well, anyway, what they did was they waited over winter and then concreted where the trails were because that's pathway. So they didn't say you're going to go on these paths and put nice bushes here. They said, where do people walk concrete levels? Emerging behaviour. Okay, so familiarity is your system intuitive? Why don't we put intuitive for little quotes? Why don't we put intuitive for little quotes? Yes? And what does Jeff Russell say about that? Yes? It's not really intuitive, it's just way of familiarity. It's just familiarity, yeah. So you've got this idea that things are intuitive that might not be to everybody and it's really about familiarity and social conditioning often times. So therefore, what might be familiar what might be intuitive to me here in this social situation may not be familiar to you lots of other people in other social situations especially when we've got global software global reach that might not be the case. Certainly, you know what, I think it's intuitive that it's probably not a very good indicator being a computer scientist and also if you know somebody with three bizarre economy tutorial knives. Those things don't make me necessarily a very good judge of this intuitive behaviour. So it's about familiarity so to do this kind of thing we need to maybe become familiar not just with different sections of our society we're building for a certain entire population. Okay, learnability are the interactions easy to learn? Well, consistency comes into this because obviously things that are consistent are obviously naturally more easy to learn things that are familiar are all the obvious to learn. Okay? But aspects of learnability are often difficult to test. How are we going to test learnability? How might we test learnability? You might you might want to know when they're going. I might have to know what you're going on. Say you're only thinking about time to task task completion time. Yes? You're going to use, like, not involved in that procedure directly in that task and get them to from a task scene, but then how long it takes them to actually learn what they have to do on their new extension. Yeah? Okay, so that's, yes? The task completion time should go dramatically down like for the first time because they're born in there how to do it for the first time when you're just doing a scroller. Yes, so that's exactly right, that generally we'd rule somebody through the same we're getting to do with a task multiple times and see whether the task completion time is different and if it isn't different, they're not learning much. Unless it's very short initially in there or we've got familiarity where we learnt it. If it's taking them the same time and that time is reasonably high, we think it's probably beyond what we would associate with familiarity then, then there's a problem. If they're not learning, it's still as complex over time. Okay. Um, roganosnes, yeah, I'm not involved with this again because we've already talked about it, so we're just talking about how the system is working on and whether it's going to fall down and again, I think this is probably, I think this is to have this either, I think it's better with the accessibility one, but I think this really ought to be some kind of development or generalised principles. Okay. So, how do I remember this? You must remember it however you like. I remember it like that. So, you can remember these principles I should wish. Now, I want to impress on you that in this exam and in this course, these principles that aren't teaching you are the ones that I think you're right. I'm going to examine you on the principles that I think you're right. But in the real world where they don't have the benefit of my extreme insight into usability, they might not think this is the case. They might want to just use some of these other possible principles that they're not just off the shelf. So, it's up to you to decide or to know these things in your outside work, but in this verify atmosphere, in this ivory tower of academia where I get to begin, at least in here is what I think to be correct. Now, in the notes, it should tell you why I think it's correct and you're free to disagree with me as much as you like. Okay, especially in the real world, you might have different views of what's being developed. Okay. Now, we've got a set of questions that I don't propose to go through completely. Okay? You should be able to see these again, like we did the accessibility ones, you should be able to see these in the notes and you should also be able to see that there's some underlying principles that justifications that go ahead of them. I just want to look at the ones that I think are a bit more, that you won't find in conventional literature if you like, ones that are just sitting there. So, stability, stability, simplicity are all there. Situational awareness, so this kind of thing is reasonably important, so orientation within the interface and within the interaction. So, what do I mean by this? Orientation. It's kind of a property of awareness. So, where are you looking? Where's your locus of attention being drawn? Jeff Ruskin's idea of this locus of attention in computers we only have one area that we're attending to. Even though in the real world we can do many in computers we seem to be very focused on one specific area. So, what we want to do is naturally draw people's attention to areas of importance. That's what headings are for in the notes. That's what images are for in the notes. So, that's what we need to think about. There are signposts, if you like. So, that asks us to orient ourselves within the actual interface and then we need to orient ourselves within the interaction just like we spoke about with, say, Amazon whereby we can see their shopping cart or other kinds of virtual systems where we can see how far or along the information we get. Now, we'll also see how this can be leveraged when we go to this as if we heard of gamification or phonology. Well, in a couple of weeks we'll be giving gamification and phonology. Yes. And we can then see how these things like this integration within the interaction can be kind of used as part of gamification because it allows us to request more information for a user profile. So, for instance, I went on to slide share and it says my user profile is 87% complete. 30% is the really juicy information. I'm not going to give them because they're going to try and sell something to me. But they want me. So, they say, oh, everybody, all your friends are 100% complete. You're at 87% complete. So, you should be 100% complete if you want to. That's the way that this gamified thing is used because you're competing against your friends who have already completed everything in there. So, this is this kind of orientation of the interaction. Okay. Is orientation navigation around and through the system easy enough? Is the error happening simple and is feedback informative? So, we don't want super complicated errors. We don't want user feedback that uses computer-georgy. We want all these things to be in user-georgy. And one way to do this is to have all of the information if we internationalize from the start, then we've got lots of different ways to describe them in the streamers. Okay, but if there's feedback and error handling. Yes? We know about internationalization and the way that we do internationalization. So, we have all these different ways of describing things to us. It's not hard coded into the system. That being the case, it's a good idea to do this because that way you can change the feedback terminology and you can change the error handling terminology as the system evolves as it gets deployed so that people say, I don't understand what that means. Right, we'll change it. What do you understand? What's the juggling we understand? It's the way we'll see in our phonology section the way that we the way that say, for instance, a cash point suddenly changes to talking to giving you its output in the company rhyming slap if you're in London. But there we are. I welcome components needed for this particular interaction visible. If they're not visible you need to signpost them. It's called preview. You need to give them some idea of preview. Of where we are. If you've only got a narrow preview of it. You can see this kind of behaviour being incomplete on web pages whereby you see that the click-through is only a couple of milliseconds long on a second one. So it clicks on a click or execute something and then immediately they return. The reason is, they haven't put in a preview that they don't know what's going to happen. They don't know where they're going because it's not signedposted to you. So they have to go there, see this is in the right place on a second one. Okay. Self-description reasonably straightforward. Aggressive disclosure something we need to think about which I think is important is does interface look over the complex? Okay, so to you does it look over the complex. When I open to say something like Microsoft Word to me it looks reasonably complex. To most users who aren't computer literate it looks reasonably complex. Yet, what do we notice about all this stuff on the what, an Andrew of Tumblr? They've stripped away lots of the stuff. It's all ready to be progressively disclosed but it's not all that because they haven't got the screen real estate. So they've progressively disclosed it specifically. So if you look at Apple pages it's far more complicated looking on Mac than it is on iPad but it doesn't have the same functionality. So because they realise that they think that an iPad is something that you need to get in too quick consumers are going to have it immediately. They don't want to know how to learn it, they just want to understand it. They want to be familiar with it. But they don't do that. They don't want to say in the same way for the Mac interface. Because it's just by the numbers. Okay. Right. Progressive disclosure with reliability, consistency we're okay with all of these. Again, field trip next week. Okay. We all know what we're doing. These will be on both sites today or tomorrow wherever I come. Any questions you've got for me remember next week we may have depending on what I feel like I feel even cooler now. Can you look with me? Okay. So you need to you can put these on on page 161 read your notes up to the self assessment I should be able to read it on page 161. Remember to have a look through the stuff on emotional and the aesthetic stuff in preparation for next week's lecture which is down there. Okay. If you want to see me come now or sit down. What do you do today? Tomorrow? Yeah. No, you do do you? Yeah. Is it a course for that? Yeah. Okay. I'm going to see you. I'm going to see you. I'm going to see you. So you're going to see those guys at the moment. Yeah. I'm going to see them on page 161. Okay. I'm going to see them on page 161. Hi. You can put these videos on anyone and you can just read them in your notebook. Yeah, so all I can do is you have to read them on page 161. Yn y cwm yw'r gwneud hynny, dyfodd o'r ddweud yma yn fwy o'r rhai. Yn y cwm yw'r gwneud hynny, dyfodd o'r ddweud hynny'n gweithio, dyfodd o'r ddweud hynny'n gweithio.