Fred van den Driessche

c is for content

Continuing the Atlas A-Z series, this article is brought to you by the letter C. C is for Content.

what is content?

That’s a very good question, especially since ‘content’ is a quite a generic word. By content we mean the core types of metadata which Atlas stores, representing real world audio/video data. Without these types Atlas would make little sense!

how is content represented?

Let’s have a look at a few of these types, shown below in a somewhat simplified type hierarchy diagram.

Simplified type hierarchy of the Atlas content types

The main type, unsurprisingly, is Content, an abstract type representing something with fields which contain metadata about audio/video data. It can contain references to topics which the content is about, a list describing cast and crew involved in its production, clips of the content and so on. Content is a subtype of Described, which defines properties such as title, description, images and, most importantly, the publisher, a.k.a source, of the metadata. Described in turn extends Identified, the root of the hierarchy, which contains the all-important id and canonical URI fields which uniquely identify the data across Atlas.

Content is extended by Container and Item which together form a hierarchical relationship: containers have many items.

Containers themselves are split into two types: Brands and Series. These two types are closely related to the US-style concepts of “Shows” and “Seasons” respectively, but a little more flexible. Brands are implicitly top-level only, whilst Series can be top-level but also subordinate to Brands. For example, according to the BBC, EastEnders is a Brand with no Series but Doctor Who is a Brand which does have Series. At the time of writing, The Escape Artist is a top-level Series, a common case when there’s only been one series so far. If there’s a second series it’s likely the two will be contained in a new Brand.

Items represent single pieces of content. Top-level, one-off, standalone shows will often be modelled as Items, currently. Unrelated, non-episodic, Items can theorectically be found inside Containers, but this is very rare. For episodic content, such as EastEnders or Doctor Who, there’s the Episode subtype of Item. There are two more Item subtypes, Film and Song, which have properties related specifically to films and songs.

Items have Broadcasts and Locations, where the Item can be found to watch on-demand, which are accessed through Versions. An Item will frequently only have one Version but some sources model subtitled and audio-described broadcasts in separate Versions.

That’s the most important parts of Atlas’ core data types covered. If you fancy a slightly less simplified version of events, have a peek below. Next time is D. D is for Denormalization.


blog comments powered by Disqus