Part 2 of code4libBC unconference of morning lightning talks on day 1.
Open Collections Overview, Paul Joseph, UBC
- open collections at UBC library
- sits on top of 4 repositories: IR (DSpace), digital images (contentdm), archival material (AtoM), open access data sets (dataverse)
- can normalize into a single model using ElasticSearch
- browse by collection, featured collections
- provide data visualizer map frequency of objects over course of time
- user can download metadata and full text
- faceted browsing
- in context highlighting of query
- great way to bring all the digital objects into a single interaction
- can stream videos, select quality of images
- researchers can register for an API key
Open Collections Architecture, Stefan Khan-Kernahan, UBC
- don’t start before you know what users want, how access, how engage, how they go about getting what they want, behaviours users exhibit with other resources that we can take advantage of
- core, common information subset for all items
- 16 core fields, 66 optional fields = metadata manual
- ingest process structure technical workflows, transforms source repository data
- for ContentDM had to map internal nicks to fields
- technical considerations: what to keep on heap vs. on disk; search performance; diacritics; multiple translations
- what it took: PHP, Python, JAVA, HTML, JavaScript (Angular, jQuery, D3, Velocity, RequireJS), toolchain (composer, grunt, bower, node, npm), XML, JSON, JSON-LD, RDF, Turtle, N-Triple (outputs, PHP scripting), elasticSearch, Nginx, PostgreSQL, PDFBox, Tika, qpdf, Adobe Media Servr, VIPS
Open Mic: What do you want from this event?
- challenge not attracting people, but recruiting content
- we’ve been more actively recruiting, which is actually against the nature of the unconference
- many expressed that they like having some talks prepared in advance to give other people ideas on what they could talk about
- unconference format might be higher risk, but force people to attend in a more collaborative mode
- a lot of the time, you end up at pure coding or pure library conferences, so like the mix of code4lib
Lunch
May do a post on breakout, but otherwise, see you tomorrow!