Sure, but then you don't want to rate a broad trend/policy by how it impacts the most high-profile cases, which by definition are mostly immune from negative consequences due to their notoriety. Move past the likes of Roald Dahl, towards less known authors, less widely spread works, and soon enough you may find works that are still culturally important, but not enough to secure their long-term survival.
> Newspapers have always been ephemeral - books have not been.
To a degree, yes. But then it's also been long recognized that local newspapers are a critical, extremely valuable window into the past. That alone justifies keeping archived copies - and I'm not talking about filling every libraries full of archival copies of major newspapers, but rather couple institutions keeping microfilms or digitized versions. As it turns out, those institutions end up assuming that the burden of preservation belongs to someone else, and this diffused responsibility leads to those "archived works" being lost forever.
> What would happen in Dahl's case is the price on the used book market would start trending upwards for the 20th century version, and at some point someone would decide to do another version to cash in and satisfy the demand.
Maybe, but I'm not sure if this holds in general, even for works of high notoriety. There may be some interest for the old edition, but if the new edition is good enough, there might be too few a people interested in the old one to make a reprint profitable.
Or at least this is how it was before. These days, the cost of storing and delivering digital copies is near-zero, print-on-demand works even for niche works and very small audiences - but there's also so much content that the work could disappear simply because nobody cared about that work in particular enough to bother keeping a copy. Bytes may be cheap, but human attention is not.