 Hello everyone and welcome to another webinar run by the Australian National Data Service, ANS. My name is Luis Martínez Uribe, I'm a research data analyst at ANS and today we have James Wilson from the University of Oxford. James took over my position after I left Oxford at the end of 2009 and since then he has been involved in a number of research data management infrastructure projects and trying to retroactively fit some strategy around these projects. He has a background in the humanities. He was manager of the Humboldt's, a comfortable humanities hub, and later he called Intuit Arts and Humanities and today he will be talking about the approach Oxford is taking to build an RDM infrastructure for the university and attempting to put that into the UK context. So he'll be talking for about 20, 30 minutes and then we'll be taking questions at the end of the talk. So you can use the question box that you will find in your go-go to webinar software in the probably your right side of the screen and you can write your questions and comments and we'll deal with them at the end of the talk. Without further ado, I'll leave you with James Wilson, all yours James. Okay, good afternoon everybody and good morning to me, so a little bit early for me at the moment, so if I say anything nonsensical then that's my excuse for doing so. So I'll be talking probably about half an hour's versus about the kind of infrastructure work we've been doing here and I might as well get straight into it. And just introduce a little bit about the first problem, I can't seem to change my slides. Here we are, here it goes. So I'll just start off by talking a little bit about the national context in the UK. Some of you may well be familiar with some of this already but this is the kind of context that we're working in. So to start off with this, it's worth pointing out that there have been several well-established subject specific data repositories in the UK for quite a number of years. The UK data archive is the example that I've noted down here. They look after mostly social sciences data with some humanities data and they've actually been going since 1967. So the concept of research data management is hardly a new one. However, over the last sort of 10 or 15 years, I think there has been a general sort of growing awareness of the importance of research data management and an awareness of the fact that the traditional subject data repositories weren't necessarily being as well used as people might have liked. And whilst they were getting some data, they weren't capturing everything and the reuse wasn't quite as great as people might expect. And I think partly in response to that, a number of the UK research councils have been introducing new and more specific requirements regarding data management planning. But it seems anecdotally from what I've been finding out by talking to people, the actual checking up on people and checking that they've actually followed their data management plans and have actually published data at the end of their research has been rather patchy. In 2004, the GISC, which is the Joint Information Systems Committee, they're basically the funding agency who are responsible for technical innovation in higher education in the UK. They kind of launched an organisation called the Digital Curation Centre who were launched in particular to kind of try and support data curation in the long term, the long term preservation and reuse. It started off as very much a library-ish kind of base thing, but over recent years becoming more and more focused on the researchers themselves. And I think that's a trend that we're seeing in the UK. In 2007, a different organisation, UConn, wrote a report for the GISC that recommended that each higher education institution should implement an institutional data management preservation and sharing policy which recommends data deposit in the appropriate open access data repository and or data centre where these exist. So in some ways, this is beginning to really mark the need for higher education institutions to become involved in this rather than simply the subject specific data repositories and individual researchers funded by the research councils here. In 2009, the GISC commenced the Managing Research Data Program, which was really a major effort to coordinate national developments, particularly institutional developments to serve research data management. And that's been responsible for funding a number of the projects that have been undertaken in the UK, including many of our own projects. Something you may well be aware of in 2009, there's an article in Nature called Data Chain from Neglect, which is really an appeal to researchers to start managing their data better and arguing that this should be an essential part of every researcher's training to know how to look after their research data. And more recently, in 2011, the Research Council's UK agreed a common set of principles on data, which they expect to be followed. And in particular, the EPSERC Research Council, the Engineering and Physical Sciences Research Council, created a specific policy of their own which is really reinvigorated attempts to re-stimulated attempts to produce better institutional research data management, because what it's done is put the emphasis much more on the institutions to actually support researchers with their research data management, rather than putting the responsibility primarily on the researchers themselves, which tended to be the case beforehand. So in 2009, when Lewis really just started working on this kind of stuff at Oxford, we produced a commitment to RDM. I'm not going to read through this in entirety. You can read it on the screen. But one thing that's worth noting here is that much of the commitment is about the latter stages of the research data management chain. So it's about the appropriate curation and preservation. It's about federated institutional data repositories, long-term management. It does indicate some things that have really informed the progress since then, such as the fact that it's needed to all be developed and supported by several departments within the university. It's not purely a library thing. But the emphasis at this point is still very much on the kind of library-side developments about the long-term preservation and curation of data. Now, it quickly became obvious to us talking to researchers and working with researchers that in some ways the kind of long-term preservation and curation of the data was only a part of what we really needed to do at Oxford. And in some ways it was a part that was already being done by the institutional repositories, the subject-specific repositories around the country. And it became apparent to us that really what we needed to do was intervene across the sort of entire research data lifecycle, beginning with actual project planning, which research counts were covering in some ways by insisting in many cases on some kind of data plan. But then we needed to work and fill in all the kind of the gaps between to make sure that the data and the metadata captured in the plan could then be passed through that as people actually created their data, they created it in such a format which would allow it to go into an institutional or subject-specific data repository and then how a suitable rediscovery and retrieval mechanism built on top of that. So this diagram is kind of our idealized vision of what we're really hoping to create with this kind of lifecycle approach. And you can see at the end there that the idea is that with the reuse, the retrieval of other people's data or the research's own data, maybe from years gone by, can then be reinserted into current research and invigorate and inspire new research, which is why you have this dashed area going back to the project planning. And all of this, all of these steps need to be underpinned really by training and support. So with this kind of idea in our minds, we kind of launched on a, well, I suppose, a program. It didn't formally begin as a program, but it's evolved into a program really as we started to put the pieces into place. So within Oxford, we really began in 2008 with an internal scoping study into data management requirements, which looked particularly at the needs of researchers. And this was something that Lewis was involved in. Then really we got started with JISC funding with their new program with the IDICSA. It's a horrible acronym, but it stands for Embedding Institutional Data Curation Services in Research. And in that project, we were really looking at preservation workflows, trying to actually build our commitment into an actual university policy. Now, it hasn't been recommended by you, Gaul, earlier, and sort of work on the kind of communications aspects, to actually communicate the basics of research data management skills and information and sources of further information to the researchers within the university. And we actually had a bit of a helping hand on the policy and kind of comms front, actually from the University of Melbourne. So we were drawing from some of the experience that the Australian universities have developed before we really knew what we were doing ourselves. So that was quite a good way to start off. And that's been followed by a number of other projects, most of which have been JISC funded, as you'll see. So the Army stands for Supporting Data Management Infrastructure for the Humanities. I hope I'm gonna get all these acronyms correct. And in that project, we started building a database as a service, which I'll come back to a bit later, and the number of training materials for researchers. The Admiral Project, which I think was a data management infrastructure for research across the life sciences, was working with people in biology and really looking at their kind of research workflows. And they started resourcing what's now becoming our kind of institutional data repository, which is called Data Bank. And this is, it's not, the work's still unfinished, but it's gradually taking shape. And this is what we'll actually be using as the sort of counterpart to our Oxford Research Archive, which is our existing archive, virtual research papers and so forth. And Data Bank will be the branch of that, if you like, which actually holds the research data to accompany it. And now as we develop it further, we're broadening it out to basically hold any kind of data which can't be held in aura. And again, I'll come back to that in a bit. The Vidas Project, that's virtual infrastructure with databases as a service, was a fairly heavily funded thing by JISC and the Higher Education Funding Council of England. And that was really to develop the databases as a service which started in Sudan into a more mature tool and to build a cloud infrastructure at Oxford so that we could set the service up as a cloud-hosted service. The Data Flow Project further developed the data stage, further developed the Data Bank data repository and also started work on a tool called Data Stage, which is kind of a counterpot to the database as a service but for unstructured data, which I shall talk about in a minute. Bringing ourselves up to the present day, we're now working on the Data Management Role out at Oxford, which is hopefully finalizing Data Bank, working more on training, reducing policy, building a data finder tool which will be effectively act as a metadata repository as a catalog of data held, not simply in Data Bank and our own local data stores at Oxford, but data held in all sorts of other places as well. And we're now starting to think about how we're actually going to integrate all the things we've got into this kind of hold, which is naturally integrated infrastructure. We're just beginning work on what's called the Ords Maturity Project. This is basically taking the database as a service to a service launch. Ords stands for the Online Research Database Service. And we're just hoping, fingers crossed, that we can get a bit more funding from Chisgen Hefti again to really commercialize the database as a service. This is something that they've been pushing for from the world go. It's part of the UK's kind of shared services agenda. Just coming this kind of coordinating role and quite keen that rather than every university simply creating their own services and investing a lot of money into something that only they're going to use, that really all these services are actually shared. And one of the ways they're looking at doing this is to get commercial companies involved or to set up kind of national spin-out companies from some of these services so that they can be provided not simply to the universities that have made them but can really be offered as a service much more broadly. And in the longer term, there's nothing really to stop these services from becoming international services as well. So most of what we're building in the moment is open source software. So people can pick up the software and customize it for their own institutions. But what Chisgen really driving for in the longer term is that these things are served on a larger scale. They're served from one place to many different institutions as a way of keeping costs down because that's something that we're having to deal with a lot throughout this is making sure that we can actually pay for everything because there isn't much for the economic conditions in the moment. There isn't much fresh money. There isn't much free money available to really do this. We have to kind of pay our own way and be very careful over the pennies. So I'm now going to take a look at the outputs, mics and then rather than getting through project by project in great detail because many of the projects have really been working on the same kind of tools. I'll actually look at the individual tools working and I hope that will add a bit of clarity. So the IXA project kind of stands a little bit because it's the first one we did because we were working with quite a narrow set of researchers who were looking at 3D heart imaging and we were really trying to understand their own requirements but then extrapolate from that the kind of requirements that researchers would have across the university at the whole. So we were in some ways trying to meet the immediate needs of that particular research group but also thinking about how to apply things more broadly. During the project we came up with a draft university data management policy, research data management policy which we worked on for quite a lot of time, got quite a lot of advice on as I've mentioned and which was then unfortunately at the end of the project rejected. So we did all this work and it didn't come to nothing but it was rejected on the basis that we were kind of obliging our researchers to do all these things and putting these obligations on departments to provide training and so forth and basically saying the researchers to use this research data management infrastructure and before we actually had the infrastructure in place. So we were basically told to go away and rethink this policy and actually develop the infrastructure first so that when we did actually publish the policy to the university as a whole people were actually in a position where they could actually follow the policy and do what they were being expected to do. Another part of the project, we've produced the research data management web portal, screenshot of which you can see on the right hand side of the screen there. This has really intended to act as the kind of single port of entry for researchers at Oxford who wanted to find out more about research data management. So the idea is you can see on that wheel that we've taken a kind of a life cycle approach here in terms of breaking up the kind of aspects of research data management they might be interested in at a given point in the kind of research cycle and each of those links through to a further menu which tells you where to go for advice, give some hints and tips, links to our own kind of training materials we've developed within the university but also to national resources and gives some quite detailed kind of information about how to come out drafting a data management plan and all those kind of things. So the idea is that this is a publicly accessible place where our accepts can go. It is open to the wider public. So if you want to take a look at it just follow the link there. So far we have had quite a lot of people listening and using a lot of the stats at hand but it seems to be pretty popular. Other things we did during IKES and we created almost as a kind of a side project really, the 3D image visualization software. We came up with a kind of a core research data metadata schema. So the kind of things, the kind of fields that we need to capture about basically all research data not simply the ones that our researchers were working with in that project we're using but a more general thing. And that was then kind of shuffled for a little bit but we're now coming back to that in the Damero project which I'll come back to. And we also look to develop a kind of a metadata addition and archiving client for the researchers. So the idea was that basically we were, well we were essentially told at the time to use our existing infrastructure in terms of our hierarchical file server and to essentially get our researchers to back up and archive things on that which they have done since Tyler Memorial but then also to add metadata with it. So we came up with a number of kind of ideas and workflows about how this might work in practice with our researchers. But there's a bit of a problem and that our HFS system was rather antique and didn't actually have any room for metadata. So whenever anyone's positive data there, there was basically, you could add, there was actually kind of one field that was free and we could either take the approach of trying to add all the metadata there or link it to metadata held somewhere else. So we took the approach of linking to metadata or elsewhere so that we could essentially change and enrich that metadata. But then there were great problems of actually keeping the metadata synchronized. The metadata held in one store synchronized with the actual data held in this rather old-fashioned backup and archive implant. So in the end, we didn't pursue that route further but that's what we investigated in that project. Okay, moving on to the actual tools we've worked on. This is one that I've been in charge of so far. This is our database as a service which will in due course be called the Oxford Research Database Service. So the database as a service is the title of the software. The Oxford Research Database Service will be the title of the service. The way this works is essentially a system for structured data or relational databases. We will in future be expanding it to deal with XML databases as well without something for the future. It begins with a researcher with a bright idea for some research. Then the idea is the researcher simply goes to a website that we support within the university where they can register their interest and they can register a project, give a bit of information about what the project's like and set aside a bit of space for the database within that project. Then basically what they get, what we deliver to them is this cloud-hosted database. Yes, centrally hosted, regularly backed up and so forth. They can import or export existing databases. So a number of the researchers we spoke to had all to access Microsoft Access databases, for instance, which they quite wanted to get online because then they could collaborate on that database with their colleagues. They could allow access to members of the public and all those things you can't really do when you've got a database on a floppy disk in a drawer, so I've got a few researchers in that kind of position. It allows people to build and structure databases from scratch or indeed if you've got an old database to restructure, that to add new tables or relationships, add columns and so forth. And that's a graphical drive and drop interface. Moving on, there's basically a generic kind of data editing interface and searching interface. The idea is that the whole piece of software is supposed to be quite easy for people to pick up and quite straightforward and work with any relational database, really. But if the generic kind of editing and searching interface isn't appropriate for your project, if you want a very smart kind of public facing front end, then that's possible so you can use the database as a service system really just at the back end and then build your own website on top of that. And certainly one of the projects that we were working with here at Oxford, the Oxford Roman Economy Project, that was very much what they had in mind because they wanted a flexible underlying database which then they could plot results on maps. So the database of the service isn't really intended to be a swanky tool for visualizing data, but rather one where you can do all the underlying basics. And the other kind of ambition really of the database as a service is that it's very easy. Once you've got the data, the metadata in this essentially stored backed up format to then simply move it into the longer term archiving. So once people no longer need the live website up and running, it can simply be packed up, moved into Data Bank or other tools and then unpacked again if needed, if other people need to come back to the data to consult it in the future. Moving on, a data stage is a very simple tool, a very simple idea anyway. It kind of complements the database as a service, but it's one that works for researchers who are really using unstructured data. So if they just have files, they can be files of any type. So they've got files and folders on their computer. Essentially it works as a kind of a drop box where people can simply host their folders remotely. So they're actually essentially hosted, again cloud hosted, where they can be backed up. But unlike drop box, it also has a number of metadata enhancements. So it's very simple to add kind of standard kind of research metadata to the kind of files and folders you're generating. So again, what can happen is you've got this idea where you can carry on working with your data whilst you're working on your research projects exactly as you would normally, but then you can choose to effectively archive, long-term archive, put it into the data bank system with its accompanying metadata when you want at the end of the project beforehand and so forth to make the whole process kind of much more straightforward than it would otherwise be. It's also very easy to change the action mission. So if you're working in a research team, you can share particular files or folders with your colleagues and keep other ones private. Data Bank, as I've mentioned, is the university's data repository to be. We're still working on it, but it's getting there these days. Basically, for the purpose of Data Bank, we define data as any research outputs that don't fit in Aura in the opposite research archive. So anything that doesn't look like a research article basically can go in here. So it can be pretty much anything really, but it's a very flexible system based on Fedora. But the idea is that people can put their data in here and they can link that data to articles in Aura. They don't have to link it to articles in Aura. It can be freestanding data, but it will all be cataloged with appropriate metadata schema and available for people to come and access, should get access, permission to be given to them. It also assigns digital object identifiers to things to make citations and so forth easier in the future. It enables embargo periods as one would expect and there's also a dark eye for research data which needs to be kept and preserved and which shouldn't be made available to the public for various ethical or other reasons. Another tool that we just started working on is the Oxford DMP online. That is data management planning online. Essentially the DMP online is a tool that the Digital Curation Center has been working on in the UK. Essentially it's an online system which is designed as a kind of a template builder for researchers who need to produce a data management plan. So the idea is you can type in or just select from a list of the research council you're bidding to and it will ask you the right questions that that research council actually wants as part of their data management plan. So instead of planning through every possible aspect of data management, it narrows it down to the ones that your research fund might actually be interested in. It enables people to create, save and submit and use their data management plans and so forth. The Oxford version of this, the original version was basically designed nationally so this is a customized version that we've been producing. But it's also effectively a bit of a critique in the original one as well because we've actually been doing quite a bit of user testing within Oxford. And the findings so far as the original version was rather technical. It's written from a kind of a sort of a library as an information management kind of perspective and background and obviously it's not really a kind of world that our research has lived in. So we've been trying to modify it and change it to really better response to what the search is looking for. The idea is that it's DataSite and Serif compatible. DataSite, I'm not sure how unknown that is. I'm, I suspect that many of you probably would have heard of it but it's establishing a kind of a core set of metadata that needs to be required in order to cite data as the name suggests. And Serif is kind of a metadata format that's worked particularly with research information systems. So the administrative people are keeping track of grant numbers, applications, principle investigators, all those kind of things to make sure that we can both input to those kind of systems and draw metadata from them as needed. And essentially we're going to make a kind of a data bank if that's probably will use a data bank system to store these data management plans. So the idea is we can store the plans alongside the actual data that projects actually come up with, making it far easier to see what people were were intending from the beginning, link it all together with information about where the funding came from and so forth, see what data was actually submitted at the end and all those kind of aspects it. So DataFinder, this is what I'm working on really at the moment. DataFinder is another piece of software which will be able to create a registry of research data. It'll have an interface for searching for data, browse and reporting it. Now this won't simply be picking up data from data bank but we'll be taking data from a number of different sources. I don't know if you can really see the architecture we've got there on the side but we're taking information from data bank. We're also taking information from the online research database service. We'll be taking information from departmental data stores. We're actually gonna be enabling people even if they've got non-electroic data to actually enter information about that data and where it's physically stored so that non-electroic data can be discovered and reused if permissions allow and the person gets in touch with the creator and so forth. It'll also be looking at some services beyond University of Oxford. So for instance, Colwiz is a web two kind of a Facebook for researchers kind of service really where research can share data and citations and so forth with each other and we'll be connecting it to that so that this data that people are sharing in this effectively commercial service outside of Oxford can also be captured and included in Data Finder. The metadata schema is based on data so I've been expanding that at the moment to make sure that we can also, we also have metadata fields that can fulfill all the requirements. All the requirements of the various research councils some of the research councils have additional requirements on top of the kind of core metadata schema that we originally came up with, I've said. And it's a hierarchical, we're building it sort of on a hierarchical structure so that you can have a sort of a departmental Data Finder for instance and then have on top of that a university Data Finder which can harvest all the information from the departmental Data Finders and then hopefully in the longer run we can actually have a sort of a national Data Finder which will be able to harvest all sorts of other Data Finders and produce kind of a national record of the various data held in various different places around the country or beyond the country. In terms of training and sports we've produced a number of different materials. We've got our main hub as I've already spoken about. We've got leaflets we've been used to hand out to new doctoral and post-doc students at the university so when we attend induction sessions we can hand this out. Basically quick guides to where you can turn to for advice. We've prepared some presentations for induction sessions which it doesn't have to be delivered by ourselves so kind of immediate team but hopefully it's simple enough to be given by people in the departments and faculties. We've got some worksheets with some hints and tips that people can read. We've actually got, we've got a system within the university called the research skills toolkit which gives information and hints and tips and software advice and all sorts of different aspects of research across all disciplines. So research data management only really forms one aspect of that. We've added a number of pages of content to that to try and fill that gap because previously there was very little research data management within Oxford. That's an example of one of the kind of, one of the kind of information pages with it. This particular one's about choosing bibliographic software. We've also prepared a couple of three hour courses with this picture of the course book you can see there which we've delivered and which we're very well received. Now all of this information is available via the Sudamie website and I think we're also now via the Damaro website as well so this is all publicly available. We have Oxford's non-branded versions so we've, we've got our Oxford branded versions you can see there but we've also done versions which intend to be picked up and used by the universities which don't have the Oxford branding. If you search the, I think it's called the JORUM, J-O-R-U-M, Learning Object Repository in the UK which is over to everyone, you'll find unbranded versions of all of these training materials in there and that's just a page within one of the, one of the exercise books. So these are intended for the three hour courses but they have kind of do-it-yourself exercises that the students who attend the course can work through while they're actually in the session. Now those are the main kind of tools you're working on. There is of course a lot of related infrastructure which wasn't specifically built as part of our research data management work which is either already existing in the university or which we're working on which kind of, we need to connect to our research data management infrastructure agency. So our HFS backup and preservation system I've mentioned already, it's been going for years, it does its job. We're working on a storage as a service system. Believe it or not, Oxford University does not actually have central storage which we can offer researches. Most universities I think do these days but until recently every department has basically had their own storage, people have stored things locally and their departmental service on their own computers but we've not actually had essentially provided storage service. We're gonna be producing one. Cole, as I've kind of mentioned, it's freely available for anyone to use. Do have a little bit, it's kind of interesting. A private cloud and virtual data center we've now got up and running. That was built as part of the ViDAS project but it's gonna be used beyond purely for research data management. It's all sorts of use that you can put too. Now we have our supercomputing center. It's not the only high performance computing center with Oxford, some of the departments of their own. Everyone tends to do their own thing in Oxford. That's gonna build it to the constitution of the university. We're developing a sort of a bigger, better research information system used by our research services. We haven't really had a single research information system before, we're working on one now and our metadata will be tied into that, as I mentioned. We've got our Oxford Research Archive for the theses and articles. We're also starting a pilot project looking at digital archiving much more broadly, not for research data but just for everything, particularly for administrative records and so forth. So to return to the life cycle I kind of showed earlier, this kind of illustrates really how the bits of infrastructure we're building kind of actually fits in to our overall vision. So the DNP online tool is very much based at the kind of an intervention of the project planning stage of any project to try and get them off in the right foot by really thinking through how they are gonna deal with their data that they're actually gonna be producing and what's gonna happen to it at the end of it. The online research database service and data stage are really services to help researchers when they're working with the active data, when they're actually conducting the research itself, when they're actually gathering the data, they're analyzing it, they're trying to produce the research outputs. So with the odds, looking at the structured data and data stage dealing with the unstructured data. The data bank system is kind of for our institutional storage and then data finder will have the role of allowing a rediscovery mechanism for the data and also a retrieval mechanism. Data finder itself won't be doing the retrieval. That will basically push people back to the source where the data that's cataloged has come from and there might well be restrictions and who can actually access the data but it'll show you what data is available and take you to the place where it can be retrieved. And underpinning all of this is the training and support the various training and support things we've already implemented and we've got a lot more work to do because most of our actual software services are still in development and we need to do a lot of documentation and training specifically on how to use those bits of software as well as the more general kind of advice and training that we've been bringing together so far. I mentioned the EXERC requirements as part of the national context at the beginning of it. They're really, they've given fresh impotence really to the whole drive in the UK to implement this kind of research data management infrastructure. Apologize, that's cool. And what they've done differently from the past is that they're really putting the emphasis on the institution rather than the researcher to have that kind of the RDM infrastructure. So whereas previously most of the research cast would have said it's the obligation of the researcher and the research group to make sure their data is well handled and becomes available afterwards and if they fail to do that, then it would be the researcher or their research group that would have problems getting funding in future. The EXERC have rechanged that, they've put the owners on the institutions and whilst previously a lot of institutions in the UK were kind of sort of monitoring developments, seeing what was going on, all of a sudden every institution has now got to have their own kind of, their own way of dealing with things in place. So everyone's got terribly excited. It's focused the minds of senior management as I mentioned there in a way that it probably worked before apart from maybe a few institutions such as Oxford, Edinburgh, Southampton, Bristol and a few others. But now it's becoming much more of a kind of a widespread mainstream concern in the UK to really get this kind of data management infrastructure up and running. They've requested a roadmap explaining how universities are actually gonna be able to do all this by 2015. So we've done ours and kind of set it off. I've only got a draft version, so I'm not gonna show it here because it would be inaccurate and not up to date. But that really is focusing minds. So we have now put ourselves a new draft policy as well. After the initial one generated in the IECSA project was kind of not approved. For reasons given earlier, we've actually got a new one. I think this is up for, it's either just been formally approved or it's up for formal approval very shortly. It's short. It's only two sides of text, albeit in a small font. It's influenced by the 10 commandments of the University of Edinburgh, which you're not familiar with already. You can have a look at it. These are basically kind of essential data management principles which state what the university should provide. A couple of quotations. University acknowledges its obligations under research funders data related policy statements. This is pretty much, I think, as a direct result of the EPSERC conditions and goes perhaps to ensure the sound systems are in place to promote best practice, including through clear policy guidance, supervision, training and support. And I have to say that this is still a little bit of a future vision because we don't really have that much in the way of data management training through supervision. I think a lot of the supervisors in the university that are looking at supervising doctoral studies and post-docs probably aren't really that expert themselves in RDM, so there's still work to do before we can really follow this policy in its full, I think, but it seems to be going through in the way it is anyway. And most previously, the initial draft policy really divided responsibilities between researchers. So researchers have given some specific responsibilities to partners given certain others. The universe to the whole certain others. It's now emphasizing this is much more of a shared kind of a shared thing. And the need to work in partnership is really being stressed much more. On the technical front, there's a minimum retention period for research data and records of three years, although many of the actual research councils requested that's much longer than that. So the EPSRC, for instance, say that research data must be held for 10 years after the last time it was accessed, which adds an extra element of difficulty on what we have to do because that means we have to really be able to monitor when the data was actually accessed as well as in keeping it for 10 years longer than that. So the three year is an absolute minimum, I think, where it's data will end up having to keep you on that. I should point out throughout all of this that if there's already good subject specific repositories, we're, as a university, not expecting to hold that data. We're expecting it to be placed in those repositories and then an entry created in Data Finder to know that people at the university have created this data and you should go to that repository to access it. So really what we're putting in Data Bank, we're putting a shared thing in there, which is kind of for all those subject disciplines which fall between the caps and there's a lot of them. Okay, actually I think I'm gonna mention this already, but this kind of highlights some of the responsibilities that we do still have, correct myself very bit, but we do still have some responsibilities which have been placed particularly on the researchers or the university as well as the shared responsibilities. So the researchers are responsible for actually managing their research data and records according to the principles. They're the ones that are actually responsible for collecting, storing it, using it and everything in such a way that we can capture the data at the end of it and planning for that ongoing custodianship, whereas the university is responsible really for making sure that the services and facilities are there that the researchers need. Okay, so that pretty much covers everything just to mention the forthcoming developments to things we're moving on to. Next things we need to really start doing is collating on the training materials and trying to embed them in the divisions and factors. Obviously we've been working so far as projects, as project teams, and whilst we can trial all our training materials, the people who work in the project teams aren't gonna be around there next year, the year after in order to really ensure that these training materials get used. We're having a big push at the moment to try and get the departments and the fact is to really take up these training materials to understand them themselves and to deploy them themselves in future years. We need to start promoting the policy to researchers. That'll be an interesting one and it'll take us a long time, I think, to really get the message through because it's quite a culture change in many disciplines. We need to really productionize the software. All the tools that we're working on, they're all kind of getting there but they're not actually kind of released as full production services yet. We'll need to add additional functionality throughout these projects. We've been interviewing researchers and trying to work with them to understand exactly what they need. We'll be covering the basics when we release the software but there's always gonna be more use cases or obscure use cases where people don't have very particular requirements about, say, who can access which bits of databases. Particularly in medicine we're finding this that they need to have very fine-grained access policies. We need to actually make sure the services are sustainable. We've done quite a bit of costing work and a lot of these services would not be able to offer free which adds an extra element of difficulty because we're expected to actually recover the costs providing these services and a lot of them aren't gonna be cheap. We need to really make sure that the metadata really can flow through that life-cycle user so that we really are gathering all the metadata and making our records as rich as we possibly can and also enabling citations and so forth and also just integrating tools, developed as well, because while I've been talking about what we're doing at Oxford, there's an awful lot of other universities who are also developing tools. We're focusing mostly on very much across-disciplinary pan-university kind of approach. A lot of other universities in the UK working on tools for particular research disciplines whereas it looks from that diagram as though we're basically gonna have covered the whole lot of the all sorts of disciplinary differences between things which means it won't work quite as smoothly and how a kind of infrastructure wouldn't be quite as full as that illustration made it look as we wanna try and integrate the work that other universities have been doing. And I think it's not the overrun my initial predictions but I think that's about it. So there's some sites there that you can refer to but I think I'm now open for questions.