The project team and representatives from all participating pilot HEIs and Data Centres convened in London for the second workshop of the UK Research Data Discovery Service on 18 February 2016. As well as taking the opportunity to meet face-to-face, rather than rely on the monthly webinars, the purpose of this workshop was to:
- Review and test the current alpha version of the Discovery Service
- Discuss requirements that still require clarification
- Finalise the metadata schema
- Discuss the scope of datasets
- Ensure participants had the opportunity to raise issues, ideas and questions
- Clarify the support required from participants for the remainder of the project
The workshop started with a presentation from Catherine Grout on how the Discovery Service fits in with the Research@Risk co-design theme from Jisc. One particular project that is closely related to this work is the Research Data Management Shared Service, providing interoperable systems for researchers and institutions to adhere to best practice throughout the RDM lifecycle. The metadata work in the Discovery Service is particular relevant to the Shared Service work and the production of a discipline neutral schema.
I gave a brief update on the status of the project, what’s been done, where we are and what needs to be done. The highlights have been on gathering user stories and requirements, development of the alpha site and harvesting datasets to populate the service. The presentation described the purpose of the workshop and gave an introduction to the group exercise. The aim of this exercise was to collect valuable feedback from participants that we as a team could act on to improve the project. The starfish exercise includes people working in groups and then putting items on sticky notes under the following categories: start, stop, do more, do less, continue. These were all related to what the project should be doing, or not doing, to support the participants and build a Discovery Service. Once all the ideas were added everyone voted on the most important. There was a wide range of suggestions, but the ones with the most votes were:
- Start – Feeding metadata from UKRDDS to data centres / repositories
- Stop – Harvesting anything but datasets, data services, software…
- Continue – Liaison with other initiatives looking at metadata for data e.g. DataCite; Developing a roadmap for the service separating out what is in version 1 and version 2.
- Do Less – Discussion on browse options
- Do More – Clarity on what the aim of the service is; Systematic user testing with researchers; Be clear about sustainability of service post project; What is the minimum useful metadata profile? Need a standard.
The exercise provided valuable feedback to the team and, although we are already working on some of the items, we will be making sure we have responded to these over the coming weeks.
To ensure as much feedback as possible was gathered there were three posters on the wall to collect Ideas, Issues and Questions throughout the day. These were reviewed at the end of the workshop to make sure we’d covered everything.
Following the group exercise, there was a session to review some of the requirements that still needed clarity. For example, what do we mean by “quality” of metadata? Is this completeness of metadata? What related datasets and information should be shown with a data record in the service? One of the issues raised with the project’s monthly webinars was that people would prefer to talk about things like requirements face-to-face and this provided the opportunity to do so. Although there were many requirements to review, I did manage to get some discussion going around the main issues and clarity on specific requirements. The conclusion of this session was that the team has enough information to implement all the requirements, if necessary, and where there is any uncertainty it can be raised with the advisory groups. Requirements will continue to be implemented and system testing will ensure functionality can be tested against these requirements. Any issues will be raised at that testing stage. As the development work is taking an agile approach we don’t have to wait until the end of the project to find out if there’s an issue. Releases will be every 2-3 weeks allowing functionality to be tested against requirements on an iterative basis.
Veerle Van den Eynden and Diana Sisu deal with engagement with the participating data centres and HEIs respectively. They have been supporting development work as metadata has been harvested to the discovery service. Over recent weeks they had been collecting feedback on the sort of support participants will require for the final months of the project. As the new metadata profile is implemented participants will require support making sure the profile is complete. Also, support on issues around licensing, extra metadata elements, working with repository suppliers will be provided.
The second half of the workshop focussed on the following three areas:
- metadata development
- system testing
- scope of datasets.
Before breaking into three groups to discuss these, a short introduction was provided for each one.
Metadata development – Dom Fripp gave an introduction to the proposed new metadata schema for the service. This had already been circulated within the team and to participants in a shared online document. The draft metadata schema should satisfy user requirements, be simple enough, learn from existing schemas rather than create a new one and be flexible enough to develop along with the service. There was agreement within the group about the core fields with discussion around specific issues such as licences, dates and harmonisation of terms. The schema will be updated to reflect the discussion and circulated to participants prior to publicising more widely.
Scope of Datasets – Veerle Van den Eynden and Diana Sisu ran this group to continue the discussion on what datasets should be included in the service. A report had been shared with participants and discussed in advisory group meetings so this session was to get further feedback prior to finalising the document. Feedback was that the service should be as inclusive as possible but differentiate as appropriate, be clear about access issues such as currently embargoed data, and although the focus is on UK data there is much stored in the UK that wasn’t created there with complex ownership issues.
System Testing – Mark Winterbottom gave a demo of the alpha system and Ade Stevenson collected feedback. The alpha site had been available online well in advance of the workshop and having functionality added on a regular basis. This session gave participants the opportunity to look at how the system works, delve into some of the more technical aspects of the service and ask technical questions of the team. Following this session, Mark collected all the actions as tickets in the project’s JIRA system, which is being used to track development work. After the workshop I reviewed these with Mark and planned out the development work for the next two months. The test site will be updated every 2-3 weeks to allow participants to test added functionality and features.
All three of these groups provided valuable feedback and produced engaging discussions. After reporting back there was time to review the ideas, issues and questions collected throughout the day before the final wrap up. Details of each part of the workshop have been provided in the links above. As the project progresses further posts will look at some of the issues dealt with in more detail, for example the metadata profile. The FAQ section of the Discovery Service will be added to and will include issues raised at the workshop.
The day ended with a thank you to all the participants and the project team for the help and support in running an engaging and productive workshop.