The debate about this issue generally has two parties: proponents of off-site storage, and proponents of having all the materials in one immediately accessible location. One of the biggest issues that's come up in this debate has to do with browsing.
The argument goes that, when you're doing research, oftentimes you'll have a general idea of the types of materials you want to use. You look up said materials and write down a few titles, a few call numbers. With this list of possible useful resources in hand, you head into the library stacks to locate said resources. Then, once you're in the general vicinity of what you want, you pull what you want from the shelves. Then, you look on the surrounding shelves to see what other items might be of relevance to your research. In other words, you browse. It's somewhat selective browsing, sure, because you're only looking in a particular area; but it's browsing nonetheless.
Now, many professors, students, and alumni are putting up quite a stink over the fact that, by moving books from the library building in order to make room for computers or *gasp* coffee shops, the library is essentially limiting the scholarly process. Browsing, they say, is key to scholarly work. And, they argue, without the ability to browse, research may become much less rounded, much less inclusive of a variety of perspectives. If you can't browse, you'll only use information that you think you need/want, and you'll never really include information that you just happen upon. Because it's hard to happen upon something that is stored at a secured facility x miles from the actual library.
I'm not entirely sure how I feel about this debate. I've heard some good arguments from both sides. For example, some people are saying that browsing itself is overrated and selective. The main argument on that point is that browsing is like shopping at a sale: while in theory you may have lots of things to choose from, in reality you're only choosing from what is still left. If the purpose of a library is to allow patrons access to materials, then patrons are probably checking things out. Thus, when a person goes to browse a shelf, they're browsing only what isn't currently checked out. Here's how Donald A. Barclay put it in the June/July 2010 issue of American Libraries:
"Because the books in highest demand are likely to be in use and, thus, off the shelf, browsing academic library shelves is the equivalent of hitting the sale tables on day three of a three-day sale" (p. 53).
Basically, there's no guarantee that you'll get what you actually want/need when you go a-browsin'.
Another interesting argument is that, because a book can only be shelved in one place despite the possibility that it's subject matter covers a variety of topics, when you're browsing you won't always be looking in the right place to find that perfect tome. Barclay's example again:
"Take the book What Are the Animals to Us?: Approaches from Science, Religion, Folklore, Literature, and Art by David Aftandilian. Virtually all U.S. academic libraries shelve this book in the zoology (QL) call numbers. Good luck to the person browsing the call numbers for religion (BL-BX), folklore (GR), literature (P-PZ), or art (N-NX), especially in a very large academic library where tens of thousands of books are shelved in each of these broad call number categories..." (p. 53).
Basically, I can see why browsing isn't generally an effective method of conducting research.
I do wonder, though, how much browsing should have to do with moving materials off-site. I'm interested to know the statistics about what percentage of library materials housed off-site are actually requested by library patrons. Research shows that the visible items are the ones that get chosen and used most frequently; what if an item is not only not visible, but not accessible until tomorrow's book delivery? Sure, off-site materials still show up in library catalogs; but do patrons find the whole off-site thing a turn-off and neglect to use these items? Disuse, to me, would be the real tell of whether or not off-site storage is a good thing.