Published on July 10th, 2012 | by Adrian Stevenson0
Timeline and workplan
1. Project Management
This work will involve monitoring the progress of the project, resolving issues and risks, and managing internal and external communications and evaluation activities. The project manager based at Mimas (Adrian Stevenson) will work with the JISC and KCL teams to finalise plans and progress the project. Evaluation criteria for the identification and selection of datasets will be developed; these criteria will be used to assess the appropriateness of fit of identified datasets from a Discovery technical and licensing point of view. The project manager will commission the interface development work, and work alongside these supplier to develop the interfaces. Blog posts on project progress, lessons learned, and particularly the alignment of the project to the Discovery technical and licensing criteria will be produced, and the project manager will consolidate and synthesis these reports for a final evaluation blog post, highlighting key challenges and opportunities in the context of Discovery. Deliverables: Finalised plan & monitoring/communications procedures; blog posts; Risk and issues log; Project evaluation and final blog posts
2. Requirements and specification development – collaborating with Kings College London
This work area will involve working collaboratively with the Kings College team to maximise potential of their scoping work, assess whether their recommendations are achievable in the timeframe, and at the same time identify quick wins in order that the API development work can progress and the final deadline for the exemplar completion is not compromised. KCL have already posited some core criteria for identifying ‘content availability.’ We will work with KCL to refine these criteria to include Discovery technical and licensing principles, and request they assist us in locating quality datasets that fulfil the minimum criteria for Discovery (see table 1). We will also work with KCL to identify baseline use cases to shape API requirements – these requirements will likely be shaped by the nature and scope of content readily available. Deliverables: Consultation, team meetings, baseline technical/metadata requirements criteria; identified datasets.
3. Data Source API development
We will work with a number of the institutions identified and prioritised in phase 1 of the project that have important content, but that don’t currently have APIs. The intention is to help them with the expertise needed to develop their own API where possible. Following an assessment of the information provided by phase 1, we will identify two to three institutions that we are most likely to be able to help, both in terms of their motivation and the resourcing available.
We will develop a set of questions to assess the feasibility of helping an institution provide an API. These will cover issues such as what systems they are currently using, and what on-going resources they have in terms of staff skills, hardware and software systems support.
Where possible, we will work with identified institutions to provide an API. The priority will be to keep any development work as quick and simple as possible. We would look to provide as sustainable an API as possible given skills and resourcing constraints. To this end we will focus on using off the shelf software such as SOLR. It is noted that we are unlikely to be able make significant changes to the underlying metadata and data structures, which may have implications for the what achievable with the aggregation API. Deliverables: APIs and documentation.
4. Develop API and Aggregation
This work will involve developing an aggregation API based on the recommendations produced in phase 1 and workpackages 2 and 3. To deliver a proof of concept for how the Discovery ‘ecosystem’ might work in practice, the team will develop an API on top of the suitable APIs identified by the Phase 1 project, and any APIs developed as part of workpackage 3. An iterative approach will be taken with specific issues tracked, analysed and where possible addressed, including: 1. Quality and consistency of metadata, and its impact on the final product; 2. Speed and results quality of the ‘meta’ API approach as compared to alternative approaches (e.g. physical harvesting). Our aim will be to develop an adaptable API as more data is integrated, or data is enriched, so that more use cases can be supported – this is dependent on the metadata quality of the contributing datasets, and lessons learned in this area will be tracked and shared. Deliverables: API and documentation
5. Interface development – commissioning and working with 3rd party suppliers
This work will involve commissioning third party suppliers to develop two interfaces using the API. We will develop a request for proposals that reflects the requirements and use cases identified through Phase 1 and Phase 2, WP 2. Early in the project, we will target suppliers with a proven track record in this area, and bids that demonstrate an innovative, iterative, and collaborative approach will be viewed favourably. Ideally the suppliers will work in parallel with the Phase 1 and 2 projects. To evaluate the interfaces, we will engage with KCL and request that their specialist groups provide feedback on any prototypes developed in the project. (see risk register for mitigating action in this area). Deliverables: front end applications using APIs (at least two); user evaluation.
|WP1: Project Management & Evaluation|
|Start-up meeting/finalise plan and procedures|
|Project blog posts|
|Final Blog post|
|WP 2: Requirements and specifications dev (KCL collab)|
|Knowledge transfer – meetings/comms|
|Agree selection criteria and approach|
|Identify datasets – 1st wave|
|Identify datasets – 2nd wave|
|WP 3: Data Sources API Development|
|Assess phase 1 datasets & select potential API candidates|
|Develop strategy for approaching institutions|
|Assess institution suitability for API development|
|Facilitate API development (where possible)|
|WP 4: Aggregation API development|
|Exploratory research – WW1 data and APIs|
|WP 5: User Interface Development|
|Requirements assessment and development|
|Commission interface developers|
|Testing and evaluation|
Image: © Imperial War Museum (Q 33814)