HyperDataset structure and terminology
The hyperdata set is organised in pyramidical/heirarchical structure that is accessed through indices. There are presently two indices: a media index which organises the data by media type and a calendar index which organises the data by date of collection. The links below the HyperDataset logo in the top right will take you to the appropriate index. Clicking the HyperDataset logo will take you the portal for the whole hyperdataset.
Each item of data is broken into an 'overview' (generally containing metadata) and one or more 'nodes'. The overview plus set of nodes taken together is referred to as the "data record." For example, the interview with the director has an overview which links to all the nodes - each node containing an extract from the interview.
Each data record has a unique accession number which is referenced in the overview. It is not necessary to know these numbers to use the hyperdataset.
Each overview contains basic metadata about the record, links to all nodes in the record and some contextual information about the record. If it is possible and meaningful to 'play' the whole record then there will be method to do so. The precise details vary according to the record's medium.
At the overview level there are 'next', 'previous' and 'index' links in the blue bar at the top. The next and previous links take you to the appropriate overview. The index takes you to the appropriate index; by default this is the media index. The calendar index can be accessed by clicking the data link in the metadata box.
At the node level the next and previous links take you to the appropriate node in the record while the 'overview' link takes you to the overview. Whenever you are in a data node you should always be able to find an overview link which will enable you to recover the context for that link.
Variations by media
Audio records. These work exactly as described above. The next and previous links are oganised by accrssion number which means there is no 'narrative' structure involved. The interviews have been broken into nodes that coincide with question and answer dyads. Occasionally they are broken down further, especially where very long answers are involved.
Video records. These work exactly as described above. The next and previous links are oganised by accrssion number which means there is no 'narrative' structure involved. The nodes are broken down into acts and scenes as specified by the camera person. They are structured by time code. e.g. 00:32:12:14 is 32 minutes, 12 seconds and 14 frames from the start of the tape. A small number of time codes are linked, meaning that they can be played in a window on the left of the screen.
Photo records. Unlike the rest of the data set, the photographs do not have distinct accession numbers - they are identified instead by an 8-digitfilename which reads ymmdd###. The y stands for year, mm for month, dd for day and ### a 3-digit number showing. The photographs themselves are organised by date of taking. Where many photographs have been taken on the same day, they have been further subdivided into albums. The overview for each photograph contains metadata exported from the camera. The node for each photograph contains the photo itself. Clicking on the photo in the node allows you to see the photograph in isolation.
Scanned Documents. These work the same way as the photographs. Unlike the photos, the scanned documents do have individual accession numbers. They are organised into groupings of similar documents.
Fieldnotes. These are organised by date of notation. Longer entries for a day have been broken into nodes. Note that Bambo Soyinka experimented with creating hypertext notes with a very small node size.