 Thank you for talking with me. It's all right. Not back to understanding. Marker Betty is a quotation in the 1932, communications with the HTML article. So, what this is all about is how can we hang our knowledge somewhere? What knowledge do we need to be able to create systems, design systems and build systems? What knowledge do we need to be able to create systems and build systems? What knowledge can we create sy'n risoedd cy Moedlu happenedaquen? Lais hyn o bachodau am bachodau am gyfnirthol wynol ni? Mae arall weithgol ond byddai hi dysgu angenпаeth a'r cael petfyniol. Does it fit? How? But not what is it that is required? What do we need? Why? This is something that we need to address. Who has done any requirements analysis? Few of you have but not many. mae nifer o'r wneud yn ystod, y cwerthu ar y cyfnodau arwag yn y pethau. Mae'r acion hwn yn ymgyrch – ac mae'r cyfnodau ffans. Mae'n gynhyrchu am y dyfodol, a rhai o'r eich cwerthu. Mae'n ceisio i'r ddweud o'r ddweud o'r cyfnodau hynny. The big list of functional things we wanted to do are non functional things we wanted to do and the non functional things are things which will more... Well, difficult to put a panel on, so we get what we'd say or needs to look nice or needs to have these kind of columns or you can do that though. We need to do a very good job with this. We mainly need to do a very good job because it's more difficult, it's less tangible because it's non functional requirements. Functional requirements, if it's an air traffic control system, Ie, we've got to the tracker plane, blah blah blah, that's kind of that, we can do that, we know what we need to do for functionality for that, it's a very easy thing to think about, but how do we need to increase the situation awareness of the controller, the radar controller for instance, is more difficult, because it's more difficult to actually decide what that means, and whether they want it, what are their problems. So, this is what it is. Now, Keto created some kind of loop to help us try and understand better the kind of cycle that we need, so this is called the awareness understanding and action cycle. So, here we have to create an awareness in ourselves of what people want, we have to really understand what people want, and then we have to do some action about it, we have to have some action. So, you can see the cycle comes in when we've got this information. So, this might be a brand new system, or, as with most systems, it might be a maintenance job. So, as we know, with software engineers you all are, 70% of all software engineers is 90 minutes. So, we've got some information to leave here, and then we need to have some purpose, what's the purpose of this information? How should it fit together? We've got some information now. Now here, we've got some choices and some actions that we can have. So, here we've got different kinds of prototypes. In our language, use of experience language, we can have different kinds of prototypes, different kinds of evaluation methodologies to understand what the user wants, what the user's satisfaction is. All very satisfied with this. And we've got some actions that we'll go through to understand if user satisfaction occurs. Now, this side we're going to look at in about the right hand side, the right hand side. We're going to look at, well, after Easter actually, because after Easter we're going to look at how to do evaluations properly, and how to understand what user satisfaction might be. What we can do stuff right now in requirements that ask this, because we can kind of, if we ask them questions, if we understand what their needs are, then we can kind of hopefully do something about those needs before there's a problem with the actual evaluation. We don't need to let it get to the evaluation stage before we have to make corrections. OK. So, here we come down to the system. In this case, Kate talked about websites, but that's not what we're going to be talking about. It's any user experience stuff. And then, from this system, we get some awareness and some understanding about user knowledge. OK. So, this is in the user knowledge domain. So, we can understand what we can become aware, what the user wants, what their knowledge requirements are, because this system will start to fail if we've created it and it's wrong. So, we can see that we don't, we can become aware of it, and then we can understand just what it is we need to do, what additional things we need to come, what additional pieces of information knowledge, interactive components, procedures, tasks. We need to actually capture the start of our model. OK. And then it goes around the cycle again, until we come to this information outcome here, which is likely to actually only happen at a version level. So, we get to a certain version, we get this information outcome, and then we'll have to start again because we're going to get a new version problem. OK. So, the process by which we kind of go through this bigger cycle is a smaller cycle that Kate talked about, because this is design, discover design. OK. So, this is something that as user experience people, you're going to do all the time. So, Penny, Penny Allen from the BBC, when she came to talk to us, spoke her out. She didn't speak about it in these terms, this discover design in use terms, but she actually did discuss this in what she was actually saying. She was saying that we discover stuff. Remember how she said, we do what we need to have. We need to become experts very quickly on a subject and we have to read lots of papers in the first few weeks so we can actually create some experiments. Do you remember that? If you don't remember that, if you look into the video, that's what she said. And then, we have to do some design. OK. We have to do some more discovery at the very end, if you like, because we're talking to people and discovering what they need. Then we have to do some design. So, remember she was talking about the design for the font system for the BBC, what the font should be like. So, she did some design for that, design for the fonts, and then she used it with people. OK, and then it actually got used in reality with designers. So, there's this cycle that we go through. Now, I would add to this cycle, as we'll see, because this is just part of the previous cycle. I would add here that we need to use and then we need to evaluate. So, we've got this evaluation or evaluation in use. OK, which also we'll see about in just after Easter. How we evaluate these things. So, here we've got these choices. We make a design, we evaluate it, we have to get some objectives out, we explore some more about what the user wants, and then we have some choices to make, and we go back into the design cycle again. So, this explore part is really what we're going to be talking about, exploring the choices here in these lectures. OK, in these next few, this next hour or one, and maybe now next week. OK. So, in the old days, what we used to do, and still, as I keep saying, in lots of companies, it's all a bit like Agile, we'll say we'll do this participatory and give all these a set of design, but nobody ever does. OK, because it's all we've got to do until the time it seems to be. So, what people used to do is we used to fetch up into a company, sit down. We used to have individuals come to us and say, this is what I want, and we used to write it down. That's the requirements. It needs to do this, it needs to do that, it needs to do that. And we've missed out a lot of information because there's a lot of implicit information that they don't tell you. There's lots of information whereby there's a political reason for me not telling you it, or there's a political reason for them to actually forget about it conveniently. OK, because it's part of the process that they do now, they just don't want to do it anymore. The managers want it, but they don't. OK. There's lots of reasons. I've gone into so many companies, what I used to do was professionally, and you go there, and generally they've got no real clue about what it is. The actual users do, OK, or want to do. The implicit nature of things is that lots of things are left unsaid. Lots of things are said because the manager might be sitting right there in the interview that you have with this person. So you can't actually talk to people freely necessarily because they're in the company line, and they're missing out a lot of information while they're doing it. So you can create the best model you like. You can create a great list of functional requirements, but it's still wrong. It might make you feel good. It might make your managers feel good. It might make their managers feel good, but it's still wrong. So this is something we try to address in this user-centred or participatory design. Well, by we have the users, or we have the users there in focus groups where managers often aren't. They come for a day, for a day. Maybe they're in, maybe we have interviews with them, maybe we have observational sessions or observational periods, which we'll get to. Whereby they're just doing their jobs every day, and you just sit there and you're taking notes, you're trying to understand what are these jobs that they're doing. Now, it's often difficult to explicitly understand that, so you need lots of different methods to be able to get the richness of information to you from them. Because as I say, there's lots of things that are said, there's lots of things that people are not necessarily truthful about directly. So, but the point of this is that at least if the users are involved in the design process, they get to say, well, we don't do this now, but we would like to do it. This functionality will help me do my job better, even though it's not there right now. You might see that emerging behaviour occurs such that there's lots of little activities outside the main or sanctioned process that go on. It will help them to get their jobs done better, where these things aren't documented. So, ES5-750 is quality procedure. Has anybody heard of ES5-750 or ISO 9000? These are quality procedures to allow procedures to be done as good as possible every time. If somebody goes off sick, they can go and get the documentation out and do that procedure with as good a quality zero people. The problem with this is that the procedure is then quite fixed, that these little things that occur outside of the actual procedure, the formal procedure, get lost, get left. And we don't get to see those things. If we look at their quality procedures and say, well, this is how they do things, then unfortunately things get missed. So, what you're looking for, as you experience people, is all the little things that help them to get their jobs done that they're not going to tell you about explicitly. Yes? Being a software engineer to ourselves, who are primarily asked to do the system functionality aspect of it, we're not becoming a super nanny for them, are we? Oh, you have distilled issue with your manager, you're not telling us. If you don't tell us, you don't get it in more extreme terms. Right, but that's not the point of this, because it's not quite, I mean that's quite a confrontational view, and that's why, you know, that kind of, the sort of, more confrontational view like that we're saying, you don't tell, you don't get, is why we have functionality creep, why, that you can write all of the actual design specifications you like, but you'll still find that after the system goes out, the company isn't happy, they want additions, they want changes. The reason why they want those changes is because we've missed it at this point. So, what we need to think about is, how can we make this occur correctly? So, you're right, we're not going to be nannies for anybody, but what we are going to do is ask the question. Okay, so, if we don't ask you the question, we're going to observe them. We're going to see what they really do, not what they say they do, what do they really do? Yes? But this sort of depends on the assumption that what people want or what people already do is what they actually need to solve their problem. That's right, and that's not necessarily true. That's not necessarily true, so that's why we have a multiple set of methods which puts the user at the centre so that they can do a number of things. You can see what they do, they can tell us about what they'd like, they can tell us their frustrations, they can tell us if there's changes or alterations or things that are dangerous maybe in the system, if they want to, if they're able to. So, what we're trying to do with participatory design is to give them the opportunity to do it where that wasn't happening before because they were very much brought to us what do you need, this functionality, go away next kind of thing. That's how normally it happens, so we'll see that in a while. It's a good point, but just because people do it now doesn't mean to say that's what they want to do in the future. It's also the case that they might want to be presented with some options because they don't know what all the options are. Obviously it was the famous four quote from Henry Ford, if you've asked what people wanted they've said a faster horse and cart when we just decided to drive the car instead. So they have to know these options and what's the moment. OK, so the first thing we need to do is poo. OK, so we need to find out who are the people. So generally, and this is generally, and you can alter this once you get experience, for all of this nothing replaces experience, OK, your experience. So once you're experienced during these user evaluations, it'll be up to you to build new ones, OK, it'll be up to you to decide which is appropriate in which context. This isn't something where we just write it down and you follow that slavishly. It's for you to decide, I'm giving you a set of tools, the combination of those tools that you use them in will give you a certain richness of result, yeah. And you might want to modify those based on the situation that you're in at the time. OK, so who are these for? So we've got actors. So we've first a specific instance of the user such as a customer manager or a sales block. So these are people, these are actors in the system. They're going to be really the users in the workplace, OK. Stakeholders. So these might be people who are not directly involved, but they still have a stake. So these might be managers who have a stake in the system. They want something from it or they want it to occur, but they're not going to use it every day. So they're not going to use it a lot probably, OK. Role, OK. So describe the person that the user will be taking. So are they here or are they, you know, somebody who's doing some kind of business analysis, some business analysis job, are they a sales person kind of thing? That kind of role, what is the role that they're doing? I'm actually into the actor quite closely into the actor. And then we've got this thing called proxy. Use to describe a person who's not really doing, who's not doing the job that they're being the actor for. So it's use to describe a person who's not a specific user who's playing that role. So they're kind of fake, if you like, a fake person, OK. Now, give me these because I want you, because these are kind of a bit more difficult to understand, you know, realising certain technology from requirements, engineering, language. What? So now, five minutes and I'm going to be strict. I want you to get into little clusters of people. Everybody should have friends if you haven't now's the time to make them. I want you to just quickly jot down the what's. So what is required from these people? What do we think will be required? That we want. That we want as user requirements people. What's the what that we want? What do we want from them? Are we not spending all four different or just a collective? Just a collective, big collective list of things that we want. What is it? I haven't got a friend that I can match, mate. I want to be on for them. Yes. Yes. Yes. Yes. Yes. Yes. Yes. Yes. Yes. Okay. Okay. I think that's five minutes. Well, it's five minutes if I'm allowed. So let's get some ideas out there. What kind of things do we want from them? What tools will make our life, what do we need? Yes. Money. Oh, dear mate. That's top of the list now. Okay. Any more? All right. Yes. Got any problems with the insistence? Probably with the systems. Any more? Maybe like the task load. In terms of how the procedure is loaded to procedures. Yes, good. The business value, there you go to choose. Any more? Yes. Strategic vision and direction. Okay. Any more? What are we all talking about? What are you doing tonight? What did you do last night? There's lots of conversation. There's all the data. Time frame. Okay. Yes. The training provided and the turnover. Training in there? And the turnover, like the how it can... How is it? The new person is going to be using it. Any more? Yes. I'm sorry. Any more? Yes. What did you work in the past? Yep, any more. Yes. There are lots of that. I just want to know. Yes, I just want to know what they do as a rock star. What do you want to do? Yes, they are all the dope ones. Yes. Where are these? They're all the dope ones. Yep. All the changes that have been made should be looked at. Changes? I hope to interact with it. Okay, any more? Honest interest. Honest interest. For a company? No. Right, okay. So, I don't even have a camera. Okay, that's good actually. I'm glad we've got that. We've got a lot of different ones. Different ones to mine. Some of them are different ones to mine. A lot of these are like, even the money one, at least it's showing the original thought, thinking differently, that this is your money, you want things like, what's the timeframe, these kind of things, priorities. So, these are kind of general planning scenarios. So, I've got things like current system. What does the current system do? Have you based it or not? So, what kind of work does it have to go on? The current documentation. So, we need the current documentation if we can, because then we can actually understand what people think should be happening, okay? Or how things should work. Or if people have put the posting notes into that documentation to say, this is a load of crap, okay? Improvements? Why improvements could be mad? Who knows? Why improvements could be mad? What changes, additions, and subtractions are required? So, what's the, you know, what kind of new approach is there? And importance. And what's important, what needs speed, what is less often used, and what is the schedule? So, what's less often used? Why does that matter? Yes? It's got a lower priority. It's got a lower priority, but also, it may have a lower priority away from the interface, okay? So, therefore, you might want to put that back in somewhere else in the interface so you don't want it right up to put her up the screen. Yes? I'm just going to say confidence. Prominence, yeah, good point. Okay. You will also need to have, it says, a plan for design, agile, participative elements, and how to have a feature creep. So, this bit at the bottom is probably most important when it comes to money, okay, and also the time frame. So, this kind of idea of feature creep is something that you really need to get a handle on. So, have you all heard of feature creep? Yeah, you all know, yeah? So, therefore, feature creep is really, as a few of you have heard it, feature creep is whereby you act your system and then you slowly start to get more and more and more and more feature requests because they can see the benefits. And then, the thing about agile development is that it can handle those features like feature creep quite well because obviously you're doing something in a very fast development cycle. Okay. But, where a new feature can be added is set in particular time. But, the point is, the requirements of the internet can ask anyone. Well, the point being is that that in the end is going to take you more time to create. Okay, which means that your total time frame, there's almost a finite time frame for the development itself always. Yeah, so, there will always be a point where the people say the client says, that's enough. I'm going to tell you now when you go away. That will always occur. So, some stuff will end up being implemented and some stuff will end up not being implemented. But, the point is that this feature can mean that that features that are required might not be in the system but features that people want might be in the system because they want them more. So, they talk about it more. So, you get this feature creep which means that that it's difficult sometimes to understand what's a really key requirement and what isn't. Okay, because people will prioritize what they want more than what is definitely required sometimes. Okay. So, if you've read what I have to make because I'm going to just give it to you but this has perhaps for understanding the thing you'll see that I actually write on page 97, there's a lot of signs of a post-it note. The whole post-it note is brilliant. It's what we should be using in most of our designs or at least for representing a lot of the designs because it's very flat to know. It doesn't scare away people if you give them a packet of post-it notes when you go to an interview session or you go to a focus group or you go to all the kind of cooperation. If you fetch up with a computer with lots of complicated software that you're trying to fill in, then that can be a problem. If you've got some official looking form that you want people to fill in with their names then that could be a problem because it can book people off and can only say certain things. If you've got a wall that's anonymous that they're not going to write their names on they're just going to have their post-it note and they're going to say what are the features we want and they can stick it in? What features don't work they can stick it on? It's anonymous. We don't know. We can't link back who's said that. So therefore they're far more likely to want to actually help you with the design by doing this. It also means that you can all stand around and put it on your wall and it also means that you can rearrange parts of the wall or parts of the post-it note to more accurately represent features that may be similar or limit or aspect which are important or less important just as you like. So it's got a very flexible feeling. The main part in the participation in designing my opinion is this aspect that they're not threatening and you'll find that when you go into these organisations often times your seems being the sort of the expert if you like that you're there and therefore because you're the expert you have more knowledge than the user and therefore they'll tell you what they think you want to hear because you're the expert you invest but that's not what you want so therefore post-it notes they're leveler they level the feel out they're very flexible and they're really easy to use so that's why you should think of them beyond things like that and of course the other positive point is that with post-it notes when you take them off the wall you start to learn and translate that into more formal language it allows you to re-read them again allows you to think about them yourself and interpret them and understand how you interpret them whereas if it's all just typed into a computer right at the time there's no chance for reflection for you about what your interpretation's been or what's been created so that's why post-it notes are excellent so how do we do this? well in general we observe we don't analyse and we then discuss so we can observe we look at what people do and note it down that's exactly what Fanny was doing in a lot of those BBC experiments and you'll see that if we get published can we talk to us that's what they will do yeah analysis so then you analyse the stuff that's produced now obviously in this analysis is your interpretation your interpretation because you know the technical field but you also have got to interpret this from what the people have said they might not know how to actually express it in technical terms properly okay however what's the negative point of this analysis fact that you're going to do this analysis yes you know what the results you're trying to get to you know what results you're trying to get to so there's almost the danger that you're going to be biasing your analysis so you're going to interpret the things according to you which means that the real part the real knowledge is lost because it's some kind of knowledge that's contained within you that's just wrapped onto this interpretive analysis that's going on I know we've got this course so we're trying to find out what these observations analysis means so this discussion aspect might be internally within your software engineering team or it might include the users reportage photo design it will include the users so you'll say well this is what I think am I right is this is are we getting on the right track here and then we design together yeah okay no coffee no coffee so there's two different types of no coffee but there's two different kinds of analysis maybe data collection data collection different kinds of data analysis so what are those two kinds there was one yes qualitative qualitative and what do those what does a qualitative mean and what does a qualitative mean qualitative is something that you actually measure how you feel, how you how you experience the normal things yeah that's right so the qualitative stuff also is maybe something that you can't measure directly so it's not just about necessarily emotions or how you feel it could be lots of different stuff but it's just something that's not measurable directly so let's look at qualitative first because in the scenario where we're gathering information from them to try and inform our design and then development we must be more interested in rich information you'll find that qualitative information is quite narrow but quite deep and you'll find that qualitative information is obviously quite broad but quite shallow the amount of information you get back from that is quite shallow initially because in quantitative information for instance you've got a questionnaire or you've decided what questions you're going to ask somebody you've decided what numbers they're going to use to rate things so you've already made some assumptions because you're asking questions which will make some assumptions because the assumptions can take within them so that's what you've got to consider now qualitative you're sitting here and saying to me about children now as I've said before we've got this thing in qualitative analysis where we do a certain kind of analysis of anything in this case to say that it's more than true in the interview but what kind of what thing do we do in qualitative analysis mainly when it comes to large bits of written work interview we can describe it and then we do something to it we code it okay but we don't code it as computer scientists as we've been discussing previously we code it in the same way that say an anthropologist codes it by looking at categories or repeated concepts so we categorise repeated concepts normally through all the different participants to see whether there's some similarity of experience or similarity of understanding okay so if we can do this we can then use those categories to maybe get a bit more information a bit more maybe qualitative information by adding a survey based on those categories but qualitative is something that you should be thinking about doing from the very start on all this requirement of dissertation okay now if you've got six months an anthropologist would be having a heart attack right now if I said you could do this in six months because they want to if you can do it in two years then they're happy normally so the six months is just because I'm changing it from years of experience because there's no way we're going to be sitting in an organisation looking at the same having a thing for two years just not having that we're not going to be looking at a thing and seeing that part there okay but the first thing is participant observation okay so this is where we're observing participants observing people doing whatever they decide they're doing at the time just participant observation came from the way from anthropology to making ethnographies whereby we'd go out to some far away country and in the worst case sit on the ground and ask people to come and talk to us and in the best case we'd just go what we do is we'd participate in their everyday life and the same is true here often it's good if you can go and do be part of a coding team and do that or be part of a user team and do the same kind of jobs that a user might be doing of your system okay such that you're integrated into that system so you learn in this case we learn by doing the job gentlemen we learn what the problems are by doing it by observing the people around us doing that job but observing their frustrations by having our frustrations ourselves and that's why it takes a long time because it takes a while for them to forget you're actually here to observe them not as a real worker if you like even though you want to do that job so it's a long term and you need to interpret what you're getting back now taking notes in this context is quite difficult and there's books about how you should take notes you can believe it there are books about how to take those notes some people suggest that the best way is to sit there with a pad and take everything copiously but that means that you're then distanced from the situation other people say it's best if you do it in half hour segments whereby you excuse yourself you need a toilet break or a fan break or whatever it is you run off and make quick notes about what's happened in that half an hour segment on a bit of paper then when you get home at night you transcribe them to some kind of electronic tool now that transcribing is called note scraping so you're scraping your notes for written handiwork handiwork and stuff to some kind of electronic system it's embedded or it's embedded which means it's embedded and invisible so this kind of work you should be invisible there's lots of work being done in Manchester and in Chicago, Bizarre where the main ethnographicals are well, two big ethnographicals anyway whereby they're mostly interesting in what we used to be called deficiency so they would go with drug users and drug users and just kind of act as part of the group gang culture, part of the gang that kind of thing so in this case it's invisible you become invisible after all what is ethnography? ethnography is where it's a discipline of anthropology and it's where you're actually creating a set of or a story based on your experiences with a set of users if you like or a set of people so therefore you've been anthropologist you want to understand more about say drug use in Manchester you go with a load of drug users in Manchester and just integrate there for two years constantly writing these notes so that at the end of that you'd write a thesis which is called an ethnography now you can also do ones using videos to see kind of video ethnographies and those kind of things and it's kind of step away from documentary work Yes? In terms of software isn't this kind of a system of observation is that any really valuable improvement of that a reinvention of the solutions? It can be, yes so it just depends about interesting for reinvention I think it's good for both and one of the reasons is I mean the likelihood is in honesty the likelihood is that you're never going to get sit to what's to do with this but you might be able to use some of the techniques in participant observation to observe and to be away and distanced in the shadows if you like from what's going on so you don't influence it too much but you still get the rich data now with regard to creating a fresh I've seen people go into organisations where it's paper based organisation or it's kind of a combination of stuff is that an iterative development if you don't make it to the computer system? Well kind of it's an iterative development of the process but it's not an iterative development to any software so I don't know I mean when we create anything unless it's absolutely fresh it's kind of an iterative development if you want to see it like that on the process at least so the key aspect of what you'll find that anthropologists talk about and ethnographers talk about is this conversations with a purpose so you have some ideas when you see somebody you observe somebody doing stuff you know you speak to music a while they might think that seems like this you've gone away you've gone for a toilet break and you've taken your notes and you think oh here's a hypothesis I think this bit of work here is about this so you didn't say I looked to them that the cooler the water cooler was saying hey that bit of work you were doing you know you start a conversation about something but it has a purpose even though it might not be completely obvious to the person who you're talking to but it has a purpose it has a purpose to you yeah so that means that you can confirm a hypothesis or some kind of understanding by not explicitly asking them because it might be that they've got more you don't want to close down the conversation directly they've got more to say than just is this the case? yes you know they've got more to say than that maybe that they'll be able to expand on it elaborate on it so that's why it's called conversations with purpose strategies to not stand out that's one of the other things that we need to consider let me just finish this slide okay so test analysis here is said to be learned by learned by doing we're learning by observing each more okay now I've put in six months here so test analysis would be really nice we have six months to do it in or we don't have six months to do it in basically often times so again it's going to be test analysis is far more short to them because you're far more directed you're interested in tasks you're interested in analysing the tasks you're not much bothered about the richness of data or information surrounding those tasks only the tasks themselves which might mean that you lose stuff you lose information but it's a trade-off you have to programmatically trade this off so it can be remount so we need to work remotely whereby we're just setting star blocks in the train centre we want to look at mobile phones and you have people used phones when they were walking and we just observe a form of people doing it okay and then as we went we created a framework from this observation so we can see the kind of tasks and categorise those kind of tasks turns out that people take five steps on average by texting between looking up and reasserting and looking down again at the text that they're doing and they follow edges so the following person who's walking straight ahead they'll follow those if they possibly can and then they take eight steps between looking up or not or they'll follow an edge okay and this was all useful because then we built a mobile platform to allow us to support that kind of behaviour on a mobile phone okay so it's notes but with a formal model of what you're actually noting down so this side it's all conversations and discussion here it's a bit more formal you've got a bit more formal you've got some more formal notes and we'll see those next time there's a lot of interpretation often because it's kind of a quantitative bit of constitiff a bit of a magic there so it's less interpretive okay I'll see you next week