I found LOC’s “What is a MARC Record, and Why Is It Important?” to be a good, straightforward piece, as a person with no real knowledge of the system. Because I know it’s a common system, controversial as it may be, and because I learn systems better with practice, I did my best to make up a record for my favorite short story collection, Good Bones and Simple Murders by Margaret Atwood, using information from my physical copy and Amazon, and then compared to the LOC’s “MARC Tags” for its entry. They were different, especially since LOC has three copies (the long entry at a glance made me worry I’d missed something significant). XML makes more sense to me (obviously; it’s words and not seemingly arbitrary numbers). It did reinforce the different fields, in any case.
The most salient point for me in Roy Tennant’s “MARC Must Die” is that cataloging has become much more thorough. It’s easier to both create records with and search more information than it really ever has been. As mentioned, records are now including information about Tables of Contents, which is highly useful, but was not seriously considered at length in drafting MARC. I am hesitant to say that MARC records are total nonsense, especially since there are clearer ways to label fields when you don’t restrict yourself to five or so characters, but kind of.