Ways to identify inferred metadata

[gautierchomel]

Due to the European Accessibility Act covering all available collections, there is currently an intense activity to analyse and enhance EPUB files.

Specifically, we know about different publishing houses, service providers and content distributors performing analysis and enhancing accessibility metadata on thousands of backlist titles. EDRLab performed and documented such an analysis on 70k titles for FeniXX (SOLIMAN: Where cultural preservation meets inclusive innovation) enhancing Readium CLI inferring accessibility metadata command capacities. This tool produces a JSON manifest that can be used to create ONIX records or write additional metadata in the EPUB file.

Despite a great attention in defining the rules, this kind of tools are subject to false positives, errors are always possible and there will certainly be more rules or changed rules in the future.

Therefore, before writing anything in the EPUB file, it is necessary to be able to identify precisely which metadata was updated by which tool, including version and date.