 They also said everyone's running ten minutes late anyway, so that's ten minutes of it gone, I think, straight off. Decision will start like ten minutes late. Yeah, see, so... Although... although... This will be the end of the talk. In this session, we're going to have four presenters. And they're going to encounter some different areas of educational resources. First of all, we'll talk about how we use OERs to support the transition of practical learners. And second of all, we'll talk about the last mentioned, the key PDF projects and the third one. And we'll be about collaborative development of OER tools and the last one is about school and data. Okay, and let's welcome our first speaker. Good morning, everyone. My name is James Brunton. I work in Dublin City University in Ireland. I'm here to talk about a project that's been ongoing for just over a year, well, a year and a half probably. The Student Success Toolbox project. This project is a funded project that is funding from Ireland's National Forum for the Enhancement of Teaching and Learning in Higher Education. And it has four project partners ourselves, DCU, especially the National Institute for Digital Learning within DCU, Maneuth University, Dundalk Institute of Technology and Sligo Institute of Technology. This project has been focused on creating digital tools that can be used by institutions or programme teams or service units to facilitate successful transition to higher education for flexible learners. In our project, we're using a broad definition of a flexible learner as any adult learner going into undergraduate study, whether that's part-time or online, off-campus, you know, distance education. The scope of this project is we're focused on supporting flexible learners through key transitions in the early stages of the study life cycle. So, thinking about study, making choices, the registration process, the first few weeks of the year, and especially that gap between maybe registering and the start of the year. That's the scope, that's as far as we're going with this project. In terms of just a very brief overview of what the project did, you know, our whole goal was to try and enhance success for flexible learners or prospective flexible learners transitioning to higher education. What we did was develop eight digital readiness tools which are being produced as OERs with CCBY licenses that we want to share with the sector. We want to share them with other people so that other people can kind of customise them for themselves and use them in their institutions. How we did this was we undertook a synthesis of the literature. We looked at the literature that was out there in this space. We also looked at the websites of some leading sort of flexible learner institutions to see what were people already using and to make a long story short, we then used that as the basis to choose what tools we were going to make in our project. As we made them, we had academics involved in the project giving feedback on the tools as they were being developed in Storyboard and Prototype, but we also brought in existing flexible learners to get their feedback on the process to try and, you know, get their thoughts on do they think this is useful, do they think this is easy to use and then if they were spotting problems or giving criticisms, we fixed those problems as we went through the process. We have a project website, studentsuccess.ie. There you get information about the project, some publications from the project and also the toolbox page. We have a Twitter account also. The toolbox page you can actually view, you can view the tools, you can sort of go through the tools yourself and see what the tools do. So we built the tools on the basis of some key principles. What we saw was the most important things that we needed to try and imbue the tools with and I'll just use that to give some examples, show you some of the tools themselves that are going through all of them. So the first one was we were very focused on trying to encourage the development of self-regulation in these sort of prospective or new flexible learners. When you boil it down to like saying one thing, what's the one thing you want to do for these people? You want to help them create a set of realistic expectations about what flexible learning is going to be about. Now this particular tool is the how much work is a tool. It's basically kind of a life domain time calculator you can use to say, right, these are the typical parts of a person's life, how much time do you spend in a week doing that and you can do it either by day or you can do it by week or you can break it down by how much sleep do I get every day because maybe at the weekends you get slightly more sleep. But time management is one of the biggest problems that flexible learners have so this tool is trying to get them to look at their lives realistically and say if you've only got two hours spare at the end of the week you're not going to get a lot of study done, you don't have to figure out some way of reducing that time. The next one was personalization. Where possible we tried to put into the tool some element of personalization where the tool would kind of take some feedback from you straight away and then augment the route that you go through the tool to give you some more specific feedback. So this is my first assignment tool, trying to give people kind of a, let people follow a narrative of what it's like to produce an assignment in higher education and that's kind of a sneaky way of trying to show people ideas about what are study skills, why might the library be important when you're studying, why might asking your tutor questions be important. But at the very start you get to kind of chart where are you now and then move on from there with a specific narrative. We thought it was very important as well not to just a lot of existing tools just use institution speak and I think that can be alienating for prospective flexible learners. So where possible you still have to give kind of neutral messages kind of the key message you're actually trying to put out there but we always tried to use sort of quotes from real flexible learners or quotes that could be from real flexible learners to try and reinforce the same messages. So here this is my computer skills again trying to key up why my computer skills and digital literacies be important for a flexible learner going into higher education. So there's kind of student quotes and then the same kind of information reinforced with just straightforward messages. And also let's say if we go back to the how much work is it calculator at the end depending on how much time you've had left over you get some advice about what that might mean for you and again that advice is given in a kind of a neutral message but then reinforced with kind of student quotes, flexible learner quotes. So always try to put things in the world of the learner in the words of learners. We tried to create the tools so that they would ideally be delivered at the point of need where they're required. So as the students are going through these very early parts of the study life cycle you know if they're thinking about study if they're at the point where they're making choices that is ideally where you want them to be really reflecting realistically on their readiness. So we have a self-readiness quiz. We have the how much work is it that kind of life domain time calculator. We've designed those to ideally drop and be used at that point and also not to give them too much information at an early point. So again if they are made a choice and they're starting to get ready then my first assignment would be a very good tool to use at that point but we're not trying to teach them study skills. If they come out of that tool knowing that study skills exist then we'd be happy. They can learn study skills once they actually get to orientation or beyond orientation. The next thing was customization. We didn't make these tools to use. Most of these tools we can't immediately even use them ourselves. We made them so that they are ready to be customized by anyone. So some of the tools have resource sections but we haven't filled those out because a particular institution might have a particular set of resources that they want to direct people to. So we've left those blank. We've left spaces in the photos where people can put in their institutional logos. Now for many of the tools we worked with a Dublin-based company FluidRock web developers to produce these. I'm a mere psychologist so I don't speak this kind of language but there is a GitHub site that people now can go to and that's the address. The tools are just there. The code is there along with information and guidelines from the web developers as to how another web developer can augment them, customize them, change them, put things in, take things out. We're also producing a kind of a guide to the tools. That's still in draft until we do some dissemination workshops in the next month. But that's going to be more on the side of what do we envision the tools being used for, when do we envision them being used for. If you wanted to use them at a different stage of the study life cycle what kind of changes do we think you'd have to make and also encompassing that within some sort of key overarching messages about what you have to do to successfully support these kinds of transitions. It is one of the tools but one of the ways that we are immediately customizing these tools is that we're taking a number of the tools and putting them together along with other content to create a five-week sort of pre-induction socialization MOOC to get ready for success MOOC which we're hoping to run for the first time probably the end of July. We're creating that MOOC on a new Moodle-based platform, Academy, and we're going to be launching that in DCU in the next month or two. To do that we're having to now go through our own first process of taking these tools and finishing them off for use within this MOOC. Just a plug, the project is up for an award at the IMS Global Learning Impact Awards in Texas at the end of May. So if anyone on the basis of what you've heard in the last 10 minutes wants to go to that link which I will tweet out when I sit down and wants to sort of rate the two bucks highly I might get a nice award for the project. So just before I finish the other thing I want to do is if anyone likes these tools are going to use the tools please get in contact with us. We're going to be customizing these tools, we're going to be following up with small-scale research projects around each of the tools because that's what the literature says we're missing good information, good follow-up especially in lots of different sites about lots of different tools about how useful these tools are so we're going to try and be doing that work. Thank you very much. We'll move this laptop without destroying it that may be too tall in the order. It's not good. My slides and my computers taking a little bit of time a little bit of extra time. Give me a sec to make sure I have them. I've prepared slides rather than doing online demo to make it really quick and to make it really technically easy don't need internet connection. Sorry? It's cool, it's just about to... Start up, we're very neatly there. So, I'm Marcel Poulter I'm a Wikimedia resident I'm a Wikimedia project contributor I've come to this conference a few times before and I usually go on about Wikipedia educational assignments. So that I don't seem like a broken record here's something different. So, on the 25th of June last year the US Supreme Court rented its judgment in Obergefell vs. Hodges and this was the judgment that legalized same-sex marriage and there's lots of celebration across social media mainly by people putting this rainbow effect on their profile pictures and nothing gets people to do this but it's kind of ephemeral I want something more meaningful I can do at the time of that Friday afternoon that celebrates what was learned so a set of civil rights for one group of people in one country has been recognized there are more battles what can I do to promote learning from that. I'm a Wikipedia contributor but the people who write about current events in Wikipedia are really good and I knew as soon as the result was announced there'd be a pretty good Wikipedia article about it there was but I saw people sharing quotes from the text of the judgment and people saying it was very beautifully written powerfully written so I could make that more accessible it's available as a PDF from the Supreme Court website but we know that official URLs change or PDFs of any self-contained thing but because it's the US government there's no copyright restriction so I could put it on Wikisauce whereas Wikipedia is the free encyclopedia Wikisauce is the free library so Wikisauce has hundreds of thousands of texts of different kinds these could be reference books sprographies, poems, plays or legal or constitutional documents so I could make a Wikisauce version of this now I suspect even if you're an open culture advocate you take the law, legal judgments that's got to be, you don't hack that you don't remix that it's got to be no derivatives but I do want to make a distinctive version of it so I copy the text and set it out to Wikisauce standards and give some context give the Wikipedia article as a link give some background as to what this case is and its significance what else I can do is add links there's a judgement site Cicero Confucius de Tocqueville about cultural ideas of marriage and it mentions other cases loving versus Virginia which if you know is the case established interracial marriage, the reality so if you know who Confucius and Cicero are and if you know what loving versus Virginia is then you've got, then that's the target audience of this article but not everybody knows that and just by the simple amazing technology of web links we can add context so it's not just a self-contained thing it assumes you know all this background information it becomes a web of knowledge linking to the soul about Alex de Tocqueville you've linked to the Wikisauce profile which tells you who he is when he wrote other people had already transcribed other relevant legislation and court judgments so I could link to that context and that would give some context of what this is I can also put the judgement in categories so because this is the same software platform the same way of working as Wikipedia so you could look this in a list of decisions on civil rights or decisions that are considered the equal protection clause of the US Constitution so it becomes fun with a bit of work I'm just one contributor to this site and other people are doing similar things it goes from something that requires an intellectual threshold to understand to something more like an educational object or something that supports an educational process something that exists in the web and you can as with Wikipedia you go to it looking at one thing and you end up with those other tabs looking at concepts you haven't heard of trying to create that process but that was just one example what I've spent a lot more time on is women's rights texts of the 17th and 18th century and the beginnings of what we now call feminism so for example in just historic texts I found paid scans of this poem by John Duncan and he's arguing that women can achieve greatness given the opportunity in the same way as men and he cites women he thinks are great poets or great writers but it's cryptic who he's referring to because he's referring to people by pseudonyms things that would be obvious to his audience in the mid 18th century but baffling now so it took a bit of research he's got explanatory notes sometimes I had to consult modern scholarship to see who he's talking about and often there wasn't information about these people on wikisauce or maybe on Wikipedia but I can make links so I created a bunch of these profiles on wikisauce for some of these authors that Duncan created up and so they give when this person moved what they're known by what they're famous for if there's a Wikipedia link to that if there's other sources of biographical information that this person gave that to context and Duncan mentions this poet doing two particularly beautiful poems and on internet archive I found a book with that poem in again in the form of page scans if something exists but it's on the web as page scans, images of pages then it's not searchable it's not fun, you haven't got the verified text of it it's kind of invisible so I transcribed that poem and that's entirely a great amount of work but that encourages another level of engagement with this Duncan poem so you can look at that in a formal and formal learning setting ask why did he choose this a particularly beautiful poem and so on oh one of the poets ended up as Susanna Duncan he married one of the poets who praised it which cast doubt on his motives in the whole exercise but so I'm kind of moving up the gloom taxonomy from spreading factual knowledge in background to endangling people to this method we could build an 18th century web these authors pay tribute to each other cite each other pay tribute to each other and these could become actual links so in one link you're taken to the thing they're talking about you're taken to background information about it I've forgotten here also includes authority files from libraries so we identify a particular person not just they're not confused with other people with a similar name the links have to work both ways otherwise it's just promoting inequality because the obscure poets will cite the grand poets and not vice versa fortunately Wikipedia and related sites have this what links here button so if I go to say on man which is one of the most famous poems of this era what comes up on what links here includes one of the poems I've transcribed which is line for light parody so you can find you can navigate through the web of information both ways and find maybe obscure stuff so that's learning I've learned a lot from that process and I hope I'm lowering the threshold of effort and resource required for someone else to run a formal course or to learn themselves about this particular topic how many minutes have I got okay very quickly I just want to point you to this there's another Wikipedia sister site called wikipedia Wikipedia is an excited that's very constraining in the type of material in the style and the referencing that you can have wiki books is a lesser known it's much more free so it's mission is educational material and it's educational material at every level so people are working on everything from preschool learning materials for counting small numbers up to postdoctoral how tos on how to do genome sequencing and biology and I'll direct you to one example this this is great this is a set of case and ideas in professional ethics this is student make stuff so students in engineering at the University of Virginia have taken a professional ethics course and they have to write a case study applying the concepts and theory from their course to a big something in the news so Edward Snowden was unprofessional he did what his paying employer wanted him not to do but in a sense he lived the values that his organisation aspires to and it's loads of things you've heard of like the phone hacking scandal or things that have been big news but just not big news for us in the west so they take a probably global approach and I recommend to you it's not all professional causing some of it's really good some of it's ok, some of it's a bit ropey and some of these case studies didn't get very far and that's good it means it's a challenging assignment and it's clear what grade the students have to get from it in the open educational movement we'd like to create huge volumes of freely reusable educational material across all subjects and levels we'd also like to involve students in the process of deciding what's knowledge and the process of publication and the process of using the search resources and summarising them in the context of this sort of community those problems solve each other and that's it oh thank you that's an ideal question I'll be taking part in the Meet the Wikipedia session over lunch but immediately after that I'm doing a sort of training session in how to put stuff on Wikisource and how we've run Wikisource events and I did a women in science Wikisource events where we published papers by female scientists who didn't have any time to find their papers you mentioned at the end to briefly the idea of students creating Wikibots could you compare that to students editing with Wikipedia pages or what are their circumstances of it too? ok so as I mentioned at the start I'm a fan of Wikipedia educational assignments Wikipedia is very constrained and for example any fact is supposed to have a citation and an inline citation and it's supposed to be an excite of pedic descriptive style of text which isn't instructional so there's a lot of kind of texts and exercises that you just can't have in Wikipedia just because of the mission of the site Wikibots is you don't have to cite everything you put in you're just required that what you're writing reflects the mainstream academic view so you couldn't put a Wikibook about your conspiracy theory so you've got to have some citations but it's much more open in the style you can write and what you can recommend that the user do thing you can kind of set assignments or you can analyse from a particular point of view like analysing in terms of the theory of professionalism or the theory of corporate error so oh yeah please do one of the restrictions within Wikipedia is that it's one item of the subject so Wikibook should you say ask your class to create a book about the subject that only exists that's a really good point so you can create it's organised by subject but it doesn't have that kind of atomism that Wikipedia has and you could have the same topic treated from different perspectives in context of different books not just the educational level but yes maybe your interest is in the theory of professionalism both Wikibooks and Wikipedia exist for their purpose of creating particular resources and don't exist primarily to support educational projects but again that's kind of an advantage and you've got to fit what the students are doing and give a raison d'etre that goes beyond the course which I thought that example was good that it's not just what we learnt in this year but it's a growing resource of case studies on this topic so it's open in just the same sense these other projects so there's 11 Wikimedia projects in total and they run on the same login system so if you've got a Wikipedia login account you've got a Wikisource and Wikibooks account and it's the same process if you have an edit button and you can make changes and theory anybody can make changes but there are cultural standards about what you can so like I said you can add links or even annotations to Wikisource text that are useful but you can't editorialise and put your own opinions in links and annotations I think that's it, thanks a lot my name is Grant Potter I work at the University of Northern British Columbia Brian works at Thompson Rivers University and we in the last few months have been working on ways that we might be able to flex some enterprise architecture to the services of open source applications as well as OER so our approach has been not simply to collaborate on OER but to collaborate on the infrastructure itself and bundle the platform with the content so over to Brian just by the way all the materials here are on etherpad documents which should allow for live editing there's a bit of a link at the top I'm about to tweet out the URL for it so if you'll notice if you agree with your devices if you hit that link at the top you'll see all the members up in the top right corner and there's a chat so if you want to chat amongst yourselves when you join for the first time you'll be prompted to log in or come in anonymously if we haven't already given you an account on our server and I don't think we have very many of you take the anonymous option so there's kind of two components to what we're about the first part that I'm going to handle I'm going to try to handle quickly because I think the part Grant's going to talk about is far more interesting but first it was how we came to be we have a problem that I think a lot of particularly smaller institutions run into where there's like one or two people that are interested in doing things outside BLE, one or two people that are interested in doing stuff that's a little bit different you probably run a WordPress installation you don't really know what you're doing you find yourself bugging the people recline hosting a lot trying to figure out why your system keeps crashing and you feel a little bit less than confident putting courses and putting really vital student stuff in these environments because you know you're kind of doing this held together by duct tape and bailing wire so what we really felt like we needed to do was a group of us we kind of realized we were all in the same position and we decided to essentially pool our efforts and we don't really have we've kind of started to become quasi-official BC campus which does open textbook stuff and other educational stuff has started to come in and play in something of a more official role we have a relationship we've seen that for infrastructure but for the most part we really do run in a very informal very direct kind of way which is I think allowed us to move really quickly in terms of practical stuff I just want to give a really brief update of a project I think I conceptually described this conference last year Jim referred to this and it was open in his keynote this morning as you can see we haven't really quite figured out what the acronym means yet we're just kind of running that script until we come up with one we like but the essential idea of the spot tools was what I was realizing was when we would get someone like Jim might come in and give a keynote and get everyone really excited like yeah I'm going to get on the open web I'm going to share my stuff and I'm going to be part of this and then maybe you get your first word press space or you get a group of professors and what do they get they get a blank templated site with no personality to it it says hello world they don't know the difference between a page and a post they don't know the difference between a tag and even when you're doing just the basic introductory workshops you're putting all this energy in time just to get people oriented to the point where they can put up their first post and what we were finding was we were just losing so many people in that space the other issue we have in British Columbians was in terms of what kind of information we can collect from our students so even having a student account inside word press where we're collecting their email address is by I think a ridiculous definition but by some definitions at some institutions in our province troublesome in terms of student privacy so the idea of these types of tools and the ones you're seeing here these were mostly built by Alan Levine the idea for each of these tools was to create the simplest possible environment for getting people stuff online that doesn't require a login so if you look at say the TRU writer space which is a very simple writing tool literally you just click a button says write and through some mojo that Alan Levine has embedded into the theme a little kind of secret one time only login is generated for this usage now obviously if you don't want just anyone to come into your space you can create a very simple code word that you just share with your class or you can enable moderation as well so that there's some sort of gatekeeping function oh I gotta say we've been having this sucker wide open since we launched it and we've yet to have a malicious post come up it's mostly just been cool things but you can see this is the word press interface so this is the first thing people see when they come to write and they're asked their title they don't have to provide their name they're given the basic word press offer an interface one of the things we learned is you can actually take a formatted word document and paste it in the copy and paste in that interface it'll maintain all the word formatting including footnotes so we've been able to very very quickly create online journals using this interface and upload header images you have the option of providing an email address if you want a secret link sent to you that would allow you to go back and edit your work later I mean that's one of the one issues we came up with this no log in no account issue was occasionally people well I want to go back and change my work now but because we didn't have a connection that was an issue so we did add that little feature you choose whatever license you want and that's the way it goes and I can show more but I guess the thing that's just been so encouraging this is an example from Grants University Biology course it's just a result this is from a course in 20th century law in Grants University it's resulted in a huge increase in adoption of open web publishing at my institution it's reduced resistance incredibly it's almost dead easy to support there's very very once you hand this over to people there's really there's no need for workshops and things like that it's a tool they intuitively know how to use right away this is a variation on this for images so if you want to create your own little mini flicker for a defined group that you run we also have it and also it's a very very again very simple posting interface that does not require accounts or logins and just one other thing before I hand things over this concept of cloning which I know there's a lot of ways of doing this and for people who do stuff with VM this may seem like old hat but again if you're at an institution and you're trying to support cool interesting open web applications and you want to avoid that abyss of the hello world what do I do next phase having a set of mostly configured mostly set up ready to run web spaces and using a tool like NS Cloner allows you to hand something over to somebody where 80% of that kind of grunt legwork to get that space to where you want it to be can be done for somebody it also would allow you to embed content within the site so we might do something like build out an ePortfolio site but with generic materials but with each section actually include the directions of what you might want to put in your publication section what you might want to put inside your personal reflection section things like that so it's been incredibly exciting but now I think I'm going to hand over to Grant because that concept of cloning applications moving things on I think goes into a much wider direction with some of the work you've been doing so some months ago maybe nine months ago the province of British Columbia had made an offer the offer was something called edgy cloud and it was targeted mostly toward enterprise applications so do you have some sort of email services or do you have some enterprise software you want to run we have free virtual machines or hand in them out you don't have to pay for them well that's really interesting I wonder how we could actually run a wide course software there's lots of folks that I know in British Columbia that are interested in using those applications but they might not necessarily have the support within their IT departments or have the skill sets within their units so we started looking at a couple different frameworks and I'm going to show you before I show you some of the frameworks we're working with and what they're capable of I'm just going to point you to that little document again you'll see I've got a little spot here says come explore with us these are servers so you can actually jump on there and start installing applications while I'm talking if you want to there's six invitations if somebody grabs an invitation you'll say sorry this invitation is invalid go to the next one go to the next one until you get one that works so I'm going to scoot over here to show you what this framework is so we have an app store with over 50 apps our apps are open source apps that people that are working on this framework are maintaining they're made available through the sandstorm.io but you're also able to take if you want to trim back the number of applications you could say let's say we have our top 10 applications that faculty are really interested in and might be useful in classes we can trim back our app store to just 10 so in this case we have it wide open so I'll just go and show you an example so here is ghost which is if you're familiar with WordPress that Brian was talking about ghost is like a walking platform really simple, lightweight so here is a course for prog rock 101 it's written in a markdown so you have all of your text you have your ability to view the live site so it will you'll see it will generate some kind of a nonsensical kind of a domain or a URL here so if I can scroll down here is all of my top 100 prog rock you must listen to and if I wanted to you'll see it down here let's say if you say well that's great that's just static text content what if you want to share that playlist what if you have some audio files some rich media that you want to share you can say well I'm going to need another app look at that so I'm going to open up that's all right I've got music too so I say I need something that I can make playlist and I need something that students together we can all generate playlists for this course so I bring up something called groove basin and I link groove basin to that course outline I think we're going to need discussion so I come back into the index I'm going to need maybe let's look for some discussions here's node bb is a discussion form and you can see how quickly it loads up so it's well I can take that I can either run it separately it's just a bunch of linked apps or I can embed some of these applications in it's just spawning here I don't think I'll quite wait for it usually takes around 10-15 seconds to start up so so that's one example of how we can use those open source applications and provision them quickly and allow faculty and instructors some real freedom and how they're going to what tools they're going to use and how they're going to share content so like here's an example of some music let's think of a directory of files here's a reading package let's say I'm not exactly sure what's in this reading package but this is an application called Davos if anyone uses Dropbox in British Columbia kind of not enthusiastic about your use of these third party file sharing tools this is a very simple file sharing tool that syncs with the own cloud mobile client as well as the desktop client you can share your reading list and you can map this to a domain so if I go if I go to publishing and I'll say if my course is at um I'm just going to use the prog prog101.ca you could have reading.prog101.ca and then make the A record and put your domain to that to that application it is really quick time to live maybe two minutes like two to three minutes so here so if I'll just go to that example so let's say if I'm going to go readings.prog .ca there's my DNS record and there's my uh so there's my text in my A record at the tree if you don't know what that means it's like what is this guy talking about my text in A record that's basic domain management stuff I have a point in URL I have a resource and I don't have a lot of resources so um now uh if I I've got my prog101 uh really nicely done I've got my all my music files I've got my reading list now I have a colleague maybe in an adjacent province or an adjacent country that wants to teach this course how do I share this with and if you try to share all that disparate content oh that might be they might not be able to manage that on their end in this case I can come up here to download backup and downloads backup that backup will contain a bunch of information as to what open source repositories are being used all the dependencies you share that with them and they can install they get all the applications the linkages in between them as well as the content so you're not just sharing the content you're sharing the infrastructure as well so uh one other so that we already have that in use we have 190 some people using it lots of positive feedback a couple different applications already in development we have some alternatives to like Google services like spreadsheet either either pad that you saw earlier we also have alternatives to analytics like using piwik easy to install same thing one click so here's another thing we've been working on sometimes there may not a time okay I'll take two minutes sometimes it hasn't been developed for the sandstone framework so we started exploring docker cloud using the bc net infrastructure so you can see here we're running 16 containers all of these some of these are databases some of these are simple like PHP apps but this I want to go quickly to show you how easy this might be let's say if I'm looking for I want to use for some reason I want to use edX I can go in here I can come in here and I can I'm not going to be able to run through the whole thing I'm running out of time but I can actually search a public index of all these different open source applications I can say I want to install on this node and I can get this URL and I click deploy and there's all kinds of other kind of different parameters and things you can add but within the span of two or three minutes you can have that deployed at your URL of your choice one more thing before I'm done we're working on a number of recipes that will allow us to share these with each other so let's say if we come up with a really good recipe for WordPress like for example Vline was talking about splots so if we want to be able to share a splot so here's a splot host this is hosted at docker cloud and edu cloud we can share this recipe we say except for you for your visual host put in your domain that you want point your a record at our server and it will be deployed in 10 minutes so if we're looking for ways hopefully to take us out of the equation so that we don't have to provide the recipes we can just share and connect to our resources and then you can serve so that's coming hopefully in the following year so just to sum it up on a level that someone as unintelligent I can understand I mean we're able to we're creating a community where we're able to share actual functionality tools and stuff ready to run the way we think about sharing an open textbook to one another that we're actually able to share the context and the tool and the environment and the recipes it incredibly speeds up our development process it dramatically drops the risk because of things like containerized hosting if one of these nodes goes nuts and someone gets hacked instead of it bringing down our whole thing we just bring down the one little node and everybody else continues on their way so it allows us to be way more open in terms of what we allow people to do and I guess at the bottom of the ethernet I ask for questions and we don't have time to really get into it but what I really would hope if any of you are interested in this stuff what I want to know is okay we got this new power how does this change how we think about what we do as educational technologists as opposed to like having a pick here come to our BLA here's our service sign up and now start doing your work inside of it how does this change do we our role in terms of how we think about linking functionalities together and applications together so this document is open and editable we have a couple questions there please ask us questions now but if something comes to you later put it in there because we're going to collect some feedback and help us move the projects forward how do you think it changes so yeah well as someone who's been kind of forced to work with the BLE you know when somebody comes to me and says well I would love to use WordPress to deliver my content but I need a a discussion space where my students can talk and I need privacy controls on that and if I you can do this stuff in WordPress but WordPress doesn't do it well and I've been through that road so now I can start thinking well okay well you can we can roll out a really great WordPress presentation template ready to roll it's kind of optimized for course type deployments like the prog rock one that that branches showed and maybe link it up to matter most which is a slack open source slack clone for people who can add multi-channel private discussions and a far more fluid interesting discussion space than anything you're going to find in any BLE that I'm aware of and with the kind of the next step we didn't get a really chance to work into it is once we start to get our authentication protocols and the ability to wrap these things together inside containers and link those things together then suddenly I've created a custom made BLE with the best in breed tools that we know of and once we've done it once like you know more grants done it more likely I can copy what he did and apply it in my space for a dimension that is something that we have working on we're testing shibboleth so we're seeing people that are offering the server space to us are also offering a Canadian access federation access to allow us to share with this not only within the province but access to our platform across the country but they're using their own institutions credentials so we don't have to manage user names and passwords they use their own institutions user names and passwords right so we hope that we currently support Google a lot too and then password just token this login thank you thank you for your presentation hey guys ok so my name is Alan first of all I'm a student at Edinburgh Uni fourth year computer science and I'm here to talk to you guys about my honors project the objective of which was to take historical documents concerning people living in and around Edinburgh about the turn of the 20th century and take them and make a searchable database out of them automatically so this is our starting point at the moment the National Library of Scotland has a lot of these post office directories which was kind of like a yellow pages of the days gone by they've got a lot of scanned pages and a lot of transcripts that have been done with optical character recognition and but the problem with this it's not really searchable what we have at the moment is either done transcripts or just straight up images so you want to go from this to something like this where you've got an actual database of people the information that the directory contains on them that's searchable and automatically derived because obviously to go through it by hand and enter this data manually into a database is incredibly time consuming so there's a few different ways to go about this one of the ways we considered was just a dictionary approach whereby we'd have a list of common words that appeared post office directories and Matlin, the surname forename, occupation etc etc but a big problem with this approach was obviously ambiguity so for instance the name Cameron could be either a forename or a surname the surname Smith could also be an occupation so that approach was kind of ruled out because it would sort of involve manual moderation the second approach that we looked at was a sort of pattern matching approach obviously this data is quite well laid out it's quite there is a structure the problem here was discrepancies in format between different directories and also within directories so as we can see here in general when a title is present the forename isn't included occupations aren't always included you know every record doesn't exactly follow the same format so we needed a way of doing this that was flexible enough to deal with these discrepancies so we decided to go for a machine learning approach so machine learning is basically an approach to classification whereby we show the computer tons to use a scientific word lots of different examples of each class we show it lots of examples of surnames lots of examples of forenames etc etc so it can learn what each of these look like and then classify unseen words so the way we do this we want to build a mathematical model of this and the way we do this is we take words and we represent them as what we call feature vectors which is essentially a list of numbers so examples are things like the word length the number of digits that was a good indicator of it being an address for example things like that there was also a dictionary lookup that we used incorporated that into the feature vector so one of the features does appear in the dictionary as a surname or a forename and so on and so forth so I've included a little sort of toy example here to sort of introduce machine learning so say we want to predict whether it will rain on any given day and we've got three features for each day we've got cloud coverage temperature and wind speed so what we do is we basically show the system a bunch of features from rainy days a bunch of features from non rainy days so it can kind of work out right it's got tons of cloud coverage low wind speed maybe a bit cold that's likely to rain whereas one with zero cloud coverage it's very unlikely it's going to rain that day so we use that approach for classifying the words and the transcription the documents into surname forename occupation title address so that was the basic gist of what we were doing so the reason why we did this as I mentioned before handled discrepancies and format very well you know the pattern approach the dictionary approach were too rigid it didn't handle ambiguity it didn't handle format discrepancies excuse me there's also the problem with the transcripts themselves the technology is quite mature now but a lot of these directories are like 150 years old so obviously the quality of the print and the paper erodes over time you know it becomes more difficult to read even for a person sometimes so a machine learning approach was also taken to correct those errors sort of by looking at we had a list of possible words and we looked at which of the words were most similar to any that didn't appear in the dictionary and used that approach to correct errors so this approach it was quite accurate we did measured accuracy by tags so we had an accuracy for the surnames accuracy for forenames so on and so forth and it was achieving between 70 and 95% accurate accuracy so like surnames and titles were extremely accurate they were up in the 90s everything else was between 70 and 85% so at the moment it sort of serves as a proof of concept because of time constraints as an honours project and other courses I couldn't do as many experiments as I might have liked to I believe we can improve this with more experiments so it's just something to bear in mind so once we have the data we're now faced with the idea of storing it and this is when the idea of open data comes in we said we needed this but is this necessarily the best way of storing it I mean this approach to storing data it's very mature, it's very well understood there's a lot of support for this kind of approach but in terms of an open data perspective it's not great particularly because here there's a very rigid schema so if we want to extend this data we're going to have a hard time doing that, we're going to have to essentially put in another field for every single record which isn't great if perhaps we have a person of particular interest that we want to add extra data to say I don't know, an author say Walter Scott was in our data and we wanted to then add his works to that we don't want to add a column of publications to everybody in the directory it just doesn't make sense so it's not exactly the most flexible way of doing it that and linking these records to data and other databases is quite difficult it's just hard it's kind of counterintuitive, this approach to databases, they're meant to exist as their own unit, they're not aware of other data but what we can do is instead of thinking about this data in terms of rows and columns and tables we can do, we can think of it in terms of subject which is the whole record, in other words the entity to which that record pertains predicates, which here would be the column titles and values which is the value in a column for a particular record and using this we can store data as a graph rather than a table so today we have record number 146 in our database I should explain here that the NLS column so the way this approach works this is resource description framework or RDF, it's been mentioned a couple of times today already so apologies if this is getting a bit familiar what we have here, that sort of shorthand for a URL so subjects and predicates are always URLs the point to either the resource that the entity in question or in the case of predicates it points to information on what values this can go from, what values this can go to, so on and so forth so there's a sort of a lot more flexible schemata for this approach it's just a case of having a URL describing whatever it is you want and then using it essentially so we've got record 146 there so we can store the surname there Smith the job they were a nurse, whoever they were and they lived on the cow gate so we can see here already if we wanted to add publications for example if we wanted to add publications it's just a case of adding another connection to another node in this case the node would be one of his publications perhaps in a database of books or whatever so the example we went for here this is what we use in the system we geo-coded all of the addresses so here we have another domain, geo could be geo names or open street map or anything like that we've got the latitude and longitude there of that address which exists in another database entirely but we can see there we can cross them with relative ease so the benefits of this project are largely to open up this data all these things aren't necessarily directly an educational resource they are kind of linked and for history students doing research and everything they don't want to get bogged down and reading through page after page after page of their old directories when there could be a searchable online resource which takes that away from them and they can concentrate on actually doing history so it provides an open source tool for exploring Scotland's history in this way and while this is limited to post office directories and in particular kind only limited to ones in Edinburgh the hope is that this can serve as a springboard for similar projects not just making this data available but also easy to access easy to search and easy to use and also I personally believe that this project is a good example of what can happen when we take the expertise of the National Library of Scotland and combine them with my own expertise as a computer scientist it demonstrates the benefit of opening your data because a lot of this data depending on the institution is quite jealously guarded for the sake of losing potential revenue streams, things like that I suppose open data doesn't really go hand in hand with monetization but it serves to demonstrate the kind of thing that can be achieved if data is made open thank you should mention the repository for the project is on my github if you want to go and check it out play around with it, that's there history students what can we do? at the moment the important thing to remember is that this project is kind of just a proof of concept at the moment and I'll finish the end product but the hope is that this will sort of help history students directly there's no reason why this same approach can be applied to other areas so it's really about creating a tool for students to use and for researchers rather than a sort of symbiotic relationship it is very much focused on creating a product for them a service, a tool thanks again thanks guys