A lot of the terminology in this can be highly confusing. Here’s a quick glossary
- Object –
- An unique (to the name space of the work space) identifier
- Bitstream that stores a digital representation of an artifact or text or metadata instance
- Compound object – basically the same as an object except that it contains multiple representaions of an object, eg a tiff image of a manuscript page and a tei encoded representation of the text
- Dataset – what scientists and social scientists call an object. Basically a set of ordered readings plus metadata defining the columns of the readings
- Collection – a group of objects collected together on a thematic basis. This will vary by discipline but examples could be date and time of data capture, geographic location of data capture, etc
Datasets are usually simple objects rather than compound objects – this is due to them being generated from a single instrument. It is possible to imagine a scenario in which two different analyses of the data are stored together.