[quink]

Yeah I don't know what the sell is there... throw away the fidelity of your data when WEMI/FRBR/BIBFRAME/semantic web is coming any ~year~ decade now, soon (lol), while re-learning everything, definitely going out to tender because your current system won't do it and shift your processes and integrations. All so you can end up halfway to DC, yeah no.

The reason libraries don't do cataloguing any longer anywhere near as much hasn't got much to do with MARC 21 being hard.

[tingletech]

The fidelity seems pretty good, at least if you convert to MARCXML and then use the XSLT from Library of Congress to generate it. IIRC it has record types for all the FRBR levels. It is also not flat like DC. It was a joy to work with from a record aggregator perspective, especially if you were generating it from MARC. You can even put the full table of contents into it. At the time that one of my colleagues wrote the "MARC Must Die" article (at least if I remember correctly) the teams working on RDA and MODS had a lot of overlap and MODS was being designed with the era's cataloging theory in mind. There was a moment in time where it seemed like a "new MARC" might go in that direction.

Having catalogers or metadata librarians write directly in MODS XML by hand never made sense (although some folks tried this), but as far as something usable to ship around I'd rather get MODS than MARC or dublin core. I really don't want to have to query a triple store to aggregate records.

Catalogers ideally would have tools that make it easy for them follow RDA/AACR2 descriptive practices without having to think about the details of MARC or MODS or linked data.

I've been out of the business for a couple of years, so I have not been following Library of Congress' BIBFRAME transition.