> Hiring librarians is not a scalable solution to fixing metadata.

I’m skeptical of this line of reasoning.

If we can’t hire librarians to label all those books, how did we ever get people to write, edit, bind, read, distribute, and finally scan them?

There are two products in Google which matter to the bottom line: core search and AdSense. If you are not one of these projects, anecdotally your requests to hire new people or get resources which cannot be requested by API get forwarded to Customer Service.

It's pretty obvious that Google just wants the raw data of the books for later projects, and the concept of making them available to the public as they have is nothing more than the thinnest of fig leaves. They're getting far, far more value from the libraries they're scanning than they are providing back to the world, and it's because of their atrocious metadata processes.

And if someone figures out how to automate the process, you'd be a fool to be "hired" by Google. That's something that should be paid for at a much higher rate than any salary.

Most people simply don't care about metadata. The average user searching for information doesn't use metadata fields to search, they just use free-form text search.

I understand the importance of good data for academics (bad metadata on electronic book archives drove me crazy when I was researching etymology for the OED), but for most people it doesn't really matter.

Thanks to all of you who replied. Interesting perspectives on things.