Detailed comparison of format registry holdings

[anjackson]

Might be useful to write up a more detailed and analytical comparison of the holdings of the different format registries, based on the Format Aggregator. e.g. what can we say about the size of the problem overall? How and why the registries differ in composition. Would make a decent iPres paper I think?

Sections:

• Overview
• Detailed comparison of registries
• Compare collection profiles with registries, including your own

To Do:

• [ ] Work out how to split the structure/analysis over the sections.
• [ ] Raise some useful errors if CSV file does not have expected columns,
• [ ] Fix forced duplication of data if you load a second CSV.
• [ ] Fill out the text so it's all more useful.
• [ ] Allow upload of DROID CSV or SiegFried YAML (load as .text then npm:js-yaml), generate extension profile from there? (Maybe a v2 thing?)
• [ ] Connect YUL/CF with other Digital Appraisal work, e.g. DPC Unconference folks (NZNL-NC, QUB-AOH, possibly BFI-LW?)
• [ ] Find ways to encourage people to share stories of the 'rare gems' in collections, to help understand the value in the long-tail of formats.

[anjackson]

This is now planned to be an iPRES poster. Focus is on both comparision and 'worst case' estimation of the size of format population.

[anjackson]

Working this up using Observable Framework, which makes nice visualisations and interactive reports. Prototyping as a 'DigiPres Workbench'. Trying to get the scope and delineation of the sections right. One minor irritation is that the venn.js library doesn't seem to cope very well if you allow dynamic changes to the set names. This makes some types of visualisation hard to implement.