binding files to one another without containers
In metadata and inline content strategies, CFAs can carry rich descriptions that explain how files relate to one another, how many files of which kinds are in pre and co sets, and so forth.
A CFA explanation is a comma-separated string with no extra spaces, consisting of 3 tokens:
nature,[clarifier],identexpr
This string is terminated by whitespace, unless identexpr contains a quoted expression (in which case, it is terminated by the close quote).
A nature characterizes the relationship between the bound file and the CFA, from the perspective of the bound file. 3 case-sensitive string values are possible. They come from the Dublin Core metadata standard:
nature | meaning |
---|---|
identifier |
There is a 1-to-1 relationship between this file and the CFA, and the identifier can be used to look up either one. The file is thus the one and only pre file in the CFA. This matches the semantics of the identifier keyword in Dublin Core. |
relation |
This file relates to something else to make its meaning clear. The file is thus co, either to a pre or to a common CFA. This matches the semantics of the relation keyword in Dublin Core. |
isPartOf |
This file is part of a multi-file pre, and the set as a whole is identified by the CFA. This matches the semantics of isPartOf in Dublin Core. |
Clarifiers are explained here. Clarifiers can be omitted, causing the first and the third commas in an explanation to appear with no intervening content.
The identexpr component of an explanation tells how this cross-file association can be distinguished from others. Normally, this is done by providing a literal identifier —a name, a UUID, a DID, a SAID, etc. In the the case of externalized SAIDs, the expression is more complex. See Identifying a CFA.