Multi-Dimensional Flatness
We had our monthly lunch club today at The White Eagle in Rhoscolyn, Ynys Môn, and in a brief conversation between Leo, Nigel and myself, Leo mentioned data lakes, which are large collections of data stored in their native format; as structured, semi-structured or binary data streams. Think database tables, Word documents, spreadsheets or audio and video files. This 'wheelie-bin' approach to storage can make access to the granular information within individual files much easier, and the whole process cheaper, to boot. As I understand it, rather than indexing, relating and applying metadata to catalogue atomic data objects for later sorting and retrieval; you rather train a system to seek, access and retrieve data for further processing using knowledge of the data structures themselves as identifiers to guide the retrieval process itself. This immediately brought to mind my original thoughts on the now sadly defunct, abandoned and totally deceased Apple Quicktime API, which proffered a somewhat different but still all-encompassing concept of data management to those who saw its true potential at the time [I have a suspicion that Apple's engineers knew this too], rather than the rather crass commercial deployment that Apple eventually delivered at the time, abandoning it completely some years ago, I'm sad to note...

Comments
Post a Comment