I, too, briefly reviewed the code, mentioned by Spencer Chase, and
represented in the Stanford github projects (presumably this is only
the code Stanford chose to make public).  I share his disappointment
at Stanford's go-it-alone choices, and I recognize that Stanford has
long had a reputation for such aloofness (not that California is alone
in that regard).

However, being a part-time coder myself (along with RF/microwave design,
atmospheric-sensing drone design, and piano roll scanning, not in any
particular order) I am somewhat sympathetic to the burn-it-all-down
impulse common among coders, which often is the "least-worst" choice
when seeking to bridge previous achievements and future goals.

I also note that there are only two contributors to the projects, which
indicates to me that someone else dumped this task on their plate, and
they knocked out some code so they could get back to what they'd rather
be doing.  The reliance on Perl and XML tells me the coders were
primarily concerned with indexing the details of curation (attribution
and provenance), and only secondarily attending to the accurate musical
reproduction of the artist's intent.  The commit comments also reveal
this lack of interest in reproduction.

It looks to me like they were mostly librarians, and only secondarily
music lovers.

Marshall Jose