According to Forbes, Apple Music Could Wreck Your iTunes Library:
At its heart, Apple Music is a simple proposition. For your monthly subscription fee, Apple will offer you access to a library of over 30 million tracks. You can listen, explore, and discover to your heart's content, and you can take that music with you wherever you go. But subscribing to Apple Music and making full use of the streaming service requires a sacrifice.
You have to hand over control of your iTunes music library to Apple and hope that Cupertino's arrogance will preserve your music collection.
[...] The issue that is upsetting many Apple users is that moment when you turn on iCloud Music for the first time and your tracks are synced to the cloud. Apple's methodology on this is not clear, but from reports and feedback from users across the internet, it appears that Apple's view of metadata and what the 'correct' track is, will take precedence over your custom edits.
The Verge's Chris Welch highlights his preference of listening to early tracks from The Beatles in mono format (just as they were recorded) rather than the automatic matching services' preferences for stereo versions. Support forums talk of collections approaching 20,00 songs becoming corrupted and full of duplicate entries, incorrect meta-data overwriting current entries, album art switched out to show the wrong albums, and more stories of personal pain.
(Score: 2) by VLM on Monday July 06 2015, @11:47AM
I've been involved in a lot of stereotypical ETL and conversion type projects and what happens is all piles of data have a certain quantity of garbage. Now under normal states there is a complicated repulsive effect that resembles the field lines of about 10 magnetic north poles pushing against each other that makes the blame game calm, unmoving, statically stable. Lets say 5% of the data is garbage, and everyone has gotten used to which data is garbage, who makes more garbage or less than average, people have processes for dealing with certain forms of garbage from certain groups, etc.
Anything that any ETL or conversion process does will dynamically change the layout and distribution of garbage. Even if you can prove that you're decreasing the garbage rate from 5% to 1%, everyone involved in the previous static system will SCREAM at the conversion team because they all had deals and understandings with each other and they have none with the conversion team so dogpile them. It can also be super disruptive to company operations because now you have a large cleanup team with nothing to do anymore, and a small undiscovered new source of garbage that'll none the less usually require more cleanup personnel and effort although some MBA claimed you'd save money in the long run, well, supposedly. You need a pretty thick skin in this line of work.
This situation is just a personal version. Some moron has completely F-ed up metadata, well, so what, he's not going to complain about himself. Someone else overwrites his wrong metadata with metadata that is a hundred times less F-ed up, he's going to whine like a baby about every little mistake in the "other" metadata.
This is all compounded by the effect that "my" stuff is a lot more important to me than a zillion people's "casual glance" opinion about the same stuff. If a hundred people don't care about something being screwed up and one does care, then its going to be screwed up. This is the well known "democratic STEM" problem where you can't figure out complicated STEM problems by doing something like running a popular election at Walmart.