Which is the organizational system used in most present day libraries? It’s my understanding that the Dewey Decimal System hasn’t been used for years, and the Library of Congress system has pretty much replaced it. Where can I have this confirmed (or denied, if I’m wrong)?
Further, what exactly is the difference between the two systems?
Dewey is still used in many public libraries, who don’t have the money to change over. College libraries are mostly Library of Congress.
Dewey uses numbers. See this for the general categories. A call number might look like 273.35
LOC uses letters and numbers. Here are their general categories. A call number might look like HT115.M38.
It’s been my experience that, within the United States, most public libraries use the DDC, while most academic libraries use the Library of Congress system. This is not a set rule - I think it’s likely that you can find public libraries using LC and academic libraries using DDC. There is also the Universal Decimal System, UDS, which is more commonly used in Europe - and I’m not sure of the difference between it and the DDC.
Just a note, that I haven’t found any specific statistics to support my statement - if none have been posted, I’ll look to see if I can find them while I’m there.
As for the difference between them - I think the entries at Wikipedia provide an acceptable overview: DDC and LC.
The big benefit of LC is that when you find one book on what you’re looking for, chances are good that there will be other books next to it that you want, too. Dewey is theoretically subject based as well, of course (I’m sure you remember from grade school the broad subject categories, right?) But it never seems to work to give me everything I’m looking for in one space, unless the library is small enough that the sections can be easily surveyed anyway. It dosen’t work well with a big, specialist library.
The benefit of Dewey… um… I guess I’m biased to LC, then. 
Dewey only really works in a smaller library, because people can’t remember a long decimal number and the classificiation system dosen’t really allow you to just remember the first bit to put you in the neighborhood in the same way as LC does. Supposedly people can only hold 7 digits in their heads at a time, and a Dewey number can be infinite. Which is fine for a small school library but not okay for a big reserach institution. LC is where it’s at, baby.
Well, I’m biased in favor of Dewey cuz I’m distantly related to him on my paternal geandmother’s side. Our common ancestor came over in 1630 which is hard to top (but then again, so what?).
My experience has been almost the opposite: The Chicago Public library system uses the Library of Congress (some libraries may have a smaller Dewey collection – for instance, the main branch does. Why? I haven’t a clue.) College was also LC. In fact, the last time I came across Dewey was in high school.
Cost is a big factor. In order to change over, you need to relabel and often move every book in the library, as well as reorganize the catalog. Public libraries are usually underfunded, so it’s easier to stick with Dewey, whereas a college can get money from alumni to pay for the job.
Now if a library is being built from scratch, LOC may be feasible, but generally public libraries don’t have the budget to make the change.
Of course Dewey and LC are both pretty arbitrary, and I assume UDS is too. I don’t know anything about UDS, but the other two both seem to work fairly well, though I prefer LC.
Of course, any arbitrary classification system breaks down for some cases. You will always find something that either doesn’t fit any of the existing categories, or fits reasonably under two or more.
The people who publish Dewey say that it is the most widely used classification in the world, used in more than 200,000 libraries.
Actually, speaking as someone who tried to do it at one point, it’s almost impossible to get an alum donor to pay for a non-capital library project. Big-money alumni want capital projects or scholarships. OTOH, colleges usually have a dedicated grant-writer, and can often pay for unusual library expenses through grants.
I wonder if the quantity of fiction plays (or played) a part in which system was chosen. I was never very impressed by the LOC fiction classification system. Separating American and British authors is kind of wacky. An average public library is probably a 50/50 split, while a college library would be more like 90/10 (WAG).
<waving hand erratically> Ooh! Ooh! I’m a catalog librarian at a university library, pick me! 
Dewey works great for small collections. Call numbers are assigned based on subject, but often there are multiple items dealing with a single subject. So in a small public library, many books in a single area can have the exact same classification number. In this case, items are further organized alpahabetically by the author’s last name. This is why I absolutely detested my shelving days as a page at the local public library - an entire range of books all beginning with the call # 369.1. groan (And yes, as was pointed out earlier, you can in theory make a Dewey number that goes on and on for nine decimal places - but this isn’t pragmatic, nor does it fit on the tiny label that we put on all the books. :D)
So say you oversee an engineering library operating on the Dewey system, and you have approximately 2,500 volumes that deal with the exact same topic - metal alloys, perhaps - so these volumes all have the same classification number. Obviously, organizing these items by author’s last name isn’t going to cut it if you are going to conduct serious, timely research. It would take forever to pull that one published by Williams in 1975, and you couldn’t find the one item that dealt with copper alloys more so than the other volumes. You’d need a way to more minutely classify the item, and distinguish it from other like items.
In comes Library of Congress classification. It’s a system that assigns a classification that is 99% unique to that item (don’t cite me on that, but there are rarely exact duplicate call numbers in LC). It begins much like Dewey, assigning a broad topic area by letter(s). It then further subdivides the topic with numbers, and possibly more letter/number combos called cutters. The last cutter (letter/number combo) decribes the author’s last name, so it’s built right into the classification for the item. The whole thing is finished off by the publication year, making it very easy to distinguish between different editions of the same work by a single author. Cool, huh? Dups under this system potentially happen when there’s two authors with the same first three letters in their last names writing about the same topic. Libraries often tweak these classifications to fit their local instances and to make everything hunky-dory in their own catalogs.
This systems is way handy and efficient, considering LC receives thousands and thousands of items to be cataloged every year. For more on LC classes themselves, check out this site.
SmackFu, the broad topic of literature is frequently broken down by period. There were no American authors in the 1500’s, and academic libraries often don’t carry a lot of popular modern fiction. That’s probably why you see a “separation” between American and British authors. I’d suggest you check this page out to compare call #'s and determine exactly why you see a difference between the way two particular works are classified. Other than that, you got me! ::shrug::
I’ve never met a Dewey library that classified their fiction at all within the Dewey numbers - they just separate children from adult and sometimes into genres and then do it by author last name. All my cataloging experience is in LC, so I don’t know if you can catalog your fiction in Dewey if you wanted to or not.
As for the way LC treats fiction, it works well in a research institution environment - where it’s useful to have the book and the books about the book together. Obviously it’s not as useful for fiction browsing. I’m sure a public library using LC could just use it for nonfiction.
From that page, English literature is PR and American literature is PS. They are further divided within, so poems, for instance, by American and British authors may be aisles apart in a large college library. I presume French translated poets would go in yet another section.
While this is a cite from fiction, I thought it might be a good one anyway. I see no reason why the author should have gotten it wrong.
Fred Lerner, “Rosetta Stone.” Originally published in Artemis I and has been reprinted in at least Year’s Best SF 5, edited by David G. Hartwell, which is my copy of it.
For a real fun time, look up a bit on the personal life of Melvil Dewey himself. One of my library school professors had an anatomically correct plush doll of him with Pants-Dropping Action.
Oh, and would-be Martha Stewarts beware - those old card catalog cabinets look like a great thrift store purchase for recipe cards and things, but be aware that in addition to being into spelling reform (note how he misspells his first name), the man was also heavy into metric reform as well. A card catalog card is not by any means the same size as your index cards. 