Embracing gesture-based computing
In the next project, Ceynowa and his colleagues sought to respond to the shift from traditional user interfaces that require a mouse and keyboard, to the new, more naturalistic interfaces that are operated with voice, touch and gestures. The Library enlisted the help of its longstanding partner, the Fraunhofer Heinrich Hertz Institute, to help them with the complexities of the technologies involved.
As this video demonstrates, the collaboration has led to the development of the 3D-BSB Explorer. Users see an image of the artefact floating above the screen, and can navigate, move, rotate and enlarge that image, simply with hand gestures in the air. They do not even need 3D glasses. "It is a much more direct and natural form of engagement and it's a big step towards the interactive library."
The metadata
To explore the metadata dimension of these developments, the Library once again worked with the Fraunhofer Heinrich Hertz Institute. The Visual Search application makes use of similarity-based image searching, a technology originally designed for the detection of copyright violations of images, which works by extracting and storing characteristics - colours; actions; texture; patterns - of the picture and making them searchable.
The solution is underpinned by a database holding some 73,000 digitised manuscripts - more than 9.5 million pages with around 2.5 million embedded images. This chimed well with the theme of OCLC's conference, dynamic data - the potential of today's huge aggregations of data to transform library services. Up to now this impressive collection of digital-image objects has not been catalogued individually, so although it is of considerable interest for researchers, it has remained hidden even after digitisation. Users can easily separate text and image searches, and can even use a slider to define the degree of similarity they are looking for. "If you set the slider to 100 per cent," explained Ceynowa, "you get very similar images back; if you slide it down, you get a broader set of results with less precise matching."
In the session, Ceynowa spelled out the game-changing implications these developments may have for the fundamental nature of metadata.
Search here is devoid of keywords because the characteristics of the pictures themselves act as metadata. So the object and the metadata are one and the same. It would be interesting to explore how OCLC's WorldCat would handle this.
Content is king but context is queen
Ceynowa threw the gauntlet out to his library peers in the audience.
Content is king, and librarians have the content. We have digitised content, born digital content, licenced content, and we have the metadata for the content. So content is king, but context is queen, and I think that we must now focus the delivery of content to the contexts of the all-pervasive internet in which our users find themselves.
Dynamic Data: a world of possibilities was held in Strasbourg and showcased exciting library innovations using aggregated datasets.
Image courtesy of Ivy Dawned via Flickr - thank you!
Sarah Bartlett is a freelance copywriter wroking mainly in the area of technologies. Her career, which has spanned systems analysis, library management and marketing includes almost ten years with library system vendor and Linked Data pioneer Talis. Sarah's website is www.bartletteditorial.com and she tweets @sarahbartlett1
.