Implement CSV-like format for URF.
URF needs a CSV-like format for efficient storage of multiple resources with similar properties. Currently the format is imagined to look roughly like the following:
The namespace declaration format is highly unstable. Similarly it unknown whether the document metadata will have separate directive lines, or if they will be described using the same columnar format but with a |<>| tag.
Currently name of the format is thought to be "Comma-separate URF" or "Columnar URF" ("CURF" or "CSURF"). The MIME content type might be text/column-urf, which actually might be interesting, as it is a play on "columnar", so perhaps "Column URF" is the best name. text/urf+csv is probably not a good choice, because the format isn't true CSV, and will deviate in significant ways from CSV that would break even the most liberal of CSV-only parsers.