Use case: Digital preservation in archives

Thanks Andreas,

Data preservation is certainly an important issue - of course it is - but I don't think we can include that as a use case as it's not specific to the CPOV. 

Obviously we're open to what other members of the WG think on this topic.

I agree with Phil that this use case (long-term stability) is somewhat fundamental to any standards effort.

Long term stability is an important facet of a standard, but that cannot - indeed MUST not - place any imposition on the method, representation or manner of storage. It MUST however ensure that it allows the data it represents to be stored in a manner which allows it to be retrieved to generate exactly the same data in all aspects - values, structure and sequence.

A good example of what I mean is the discussion over whether events should have some form of sequencing via reference to previous/next node, or sequencing via an ordinal. The use of an ordinal does not guarantee that a stored and retrieved model would retain the same order, or that a full sequence can be generated from an arbirtrarily selected node. This is because an ordinal does not specify maximum or minimum values, start points, direction, etc.

The use of a previous/next reference(s) does allow the sequencing to be reproduced faithfully after storage in, and retrieval from, any arbitrary storage mechanism in any non-lossy format, and can be fully reproduced from an arbitrary node.

We should aim to ensure this 'Repeatability through Storage' (can't think of a better name at the moment) mantra is considered during CPOV defining.

Indeed an general issue for any data standardisation and management activity, but rather out of scope of the CPOV. 

It was agreed in the meeting of 2016-03-09 that, although clearly important, digital preservation is out of scope for the current work.