Their justification was that TAOCP had never been checked out. Which is funny because the year prior when I tried to check out a volume, they told me it was a reference book and could not be checked out.
I'm still completely confused about how that decisions was actually made, and why the books weren't given away or sold.
TACOP is extremely dense and put me to sleep every time I tried to read it. I recently donated my copy of 0-4 to my work library, where nobody was interested in reading it.
The fact that we both have opposite views of your example books shows that this is a difficult process for librarians.
That said, I never worked through the remaining chapters of the series, and likely never will. It's a reference book. Read the sections as they pertain to you. For instance, TAOCP was my introduction to sorting networks. This led to direct, measurable, improvements in a few embedded programs (where consistent, deterministic behavior, and optimal behavior, were strongly desired).
I mean, few people ever read a dictionary. I still have one on my desk at home and at the office. This isn't substantially different (though I do recommend skimming it so you are aware of what resources it contains for later reference).
Ideally one or two shallow-SMEs in each field would be maintained over a region, but that will never happen.
So you end-up with librarians making retention decisions based on what publications tell them are the "most useful" books in a topic area and trying to correlate that with check-out frequency. But those topic areas are coarse-grained like 'history of aviation' or 'web programming'. It's a crap-shoot and unfortunately a lot of good books are weeded instead of the dross.
My cheap solution: non-fiction is retained until a updated edition is purchased ( or forever if that doesn't occur ), non-curriculum fiction is weeded after 50 loans regardless of replacement. Curriculum fiction such as Shakespeare probably shouldn't be weeded, just replaced.
One of my mates was a DJ. He had the rule of 100. He only owned 100 records. Whenever he bought a new one, he had to purge another, to stay at 100. Some tough choices were made.
My local library auctions their books. The books I donate get added to the pile. Extras are probably bought in bulk by Powell's or thriftbooks.com
Another option might be to try putting them on the "library book sale" shelf for ultra cheap... if that's legal.
A lot of people don't realize the detailed process that goes into deciding to weed a book. When was it last checked out? How does it compare to other materials in the collection? How many copies of this book are available regionally, nationally? What's the monetary value of the book? Is the subject matter outdated, and if so, does it still have value? It's a very complicated process and unfortunately libraries simply aren't able to keep every book, nor should they!
As a compromise I've seen some libraries offer weeded books for very cheap or even give them away for free, which seems like a good way to please both sides.