As a Linked Data user, , I want to model time based entities in the DAAN/B&G schema

[mwigham]

Recognised speakers, faces, places, text etc. are modelled in as logtrack items in DAAN. We need to find a way of modelling these in linked data that makes them usable for linked data users. We will need to consider both how people will use the linked data, and what is practical given the linked data technology (e.g. complex text search is probably better suited to the Media Suite than to linked data).

Depending on how diverse the potential uses and the data types are, we may end up splitting this story over multiple issues, one per recognised data type (e.g. one for speakers, one for ASR...)

Tasks:

• [ ] Discuss potential uses of this sort of information, and determine which uses are appropriate to Linked data
• [ ] Model the logtrack item information as Linked Data in such a way that it supports the selected uses of the data.

Acceptance criteria:

• [ ] The model supports the chosen uses of the recognised items in linked data
• [ ] The model is correct RDF

[jblom]

@mwigham a bit too vague and quite far from must-have at this moment. Put it in the icebox, so we can work out the criteria better at a later time (if you agree)

[mwigham]

This issue related to the bg:transcripts metadata field. We will need to re-evaluate with the DAAN metadata to see if this issue is still relevant. It depends on how the enrichments (both old and new) are modelled in the logtrack items.

[jblom]

@mwigham I would rephrase this issue as:

as a data engineer, I want to model time based entities in the DAAN/B&G schema.

I think it's much easier to understand what needs to be modelled; where these entities are from does not matter

[mwigham]

@jblom I agree, I've changed it. I have also updated the description to fit with the way the data is modelled in DAAN, as we will no longer be doing this for the iMMIx data.

[jblom]

@mwigham thanks, much clearer now!