A central archive for personal media files.
- Data in the the repository is immutable
- Data is stored in a content-addressable way leading to automatic deduplication
- Metadata for the stored data can be stored in the repository or next to it in a simple enough format to allow access to metadata in the future
- Metadata can be either entered manually (ownership data, comments, ratings, etc.) or be extracted by pluggable metadata extractors (file metadata like EXIF, basic file information, contents of file containers, keyframes of videos, text content of PDFs, any summary of file you might want to query)
- Metadata Processes can enumerate all the existing metadata and do any kind of aggregation of the total data and provide access to it
- Complete processing might take a long time, so it needs to be incremental