> Sending books to be pulped is always sad, but the truth is that there are just too many books.
Often it's not obsolete books that are destroyed, where there might be 1,000 other copies out there, but one-of-a-kind archival records:
>> The core of the book recounts Baker's attempt, in 1999, to persuade the British Library not to junk more than 2,000 bound volumes of American newspapers - the last remaining copies in the world - including a complete run of the Chicago Tribune from 1888 to 1958 and hundreds of editions of Joseph Pulitzer's ground-breaking colour broadsheet of the 1890s, the New York World.
> A researcher doesn't gain a great deal of benefit from seeing an original paper copy; in most cases, they'd prefer a digital copy. Preserving paper is inherently more expensive than preserving microfiche or digital scans.
Maybe now, but preservation isn't about now, it's about the distant future. Paper is one of the few storage mediums that has proven longevity and also requires no special reading hardware. Microfiche is probably OK (there was an article recently that stated it could lat 500 years), but I'd dispute strongly that "preserving paper is inherently more expensive than preserving ... digital scans." Digital data requires massive amounts of infrastructure to be usable (hardware, software, storage media, and reading hardware), and the rapid pace of change of that infrastructure means that enormous, continuous efforts must be made to keep it readable in the future. Seal a newspaper up in a wall and it'll still be readable in 100 years. If you stop your digital preservation efforts for that long, you've lost everything.
https://en.wikipedia.org/wiki/Digital_preservation