Cat. & Reference: Serials vs. Monographs
The serial / monograph distinction is one of the fundamental dichotomies in the cataloging world. Serials are items that are published in an unending series, such as magazines and newspapers. "Periodicals" would be a pretty good synonym, although the term "periodical" implies that the item is published at regular intervals, which actually doesn't have to be the case with all serials. Monographs are items like books, which stand alone and aren't published serially. (The word "monograph" can also refer to a specific type of scholarly treatise, but librarians use it in a much more general sense.)
In some ways, serial cataloging is easier than monographic cataloging. In some ways, it's harder. Item per item, it's easier. This is because you only have to have one record for an entire serial, not one record for every volume or issue of a serial. So, when we get a new issue of The New Yorker in every week, we don't have to create an entirely new record, we just have to update the holdings area to include the latest issue. In contrast, every time we get a new book / monograph, we have to create or copy a completely new record, unless the book happens to be a second copy of a book we already have or a missing volume from a multivolume set. Really, serials cataloging should be the easiest thing out there, except for a troubling property of serials: They're ALIVE!
I don't quite mean this in the B horror movie sense (although there are some disturbing parallels). Part of the definition of a serial is that they're not ever supposed to end. This, of course, isn't at all reflected in reality, where most serial publications meet quiet deaths due to lack of funds or interest. However, if one originally intended to publish new volumes or issues for as long as possible, then it counts as a serial. (Incidentally, the span of publication dates may be a good indicator for determining if something is a serial or a monographic set, but it's not fool proof. I've cataloged multivolume sets which had publication dates spanning over 100 years, and serials which consisted of only one issue.)
If I'm cataloging the second edition of a book, and the title's changed since the first edition, it's not a big deal. I'll probably, as a matter of courtesy, include an original version note that indicates what the former title was. In rare circumstances, I might create a uniform title for the earlier edition. What I won't do is go back and change anything about the record for the earlier edition; it was meant to describe the book at the time it was published, and it still does.
However, serial records are meant to describe all issues of a particular serial, which is both a blessing and a curse. The curse comes in to play when anything about the serial changes, such as the title, editor, issuing body, frequency of publication, etc. Whenever this happens, you have to go back to the old record, update the information, and put the old information in a note. Like I said, serials are "alive," which means that you're not just recording information about the publication at one moment in time, but all the changes that have been made over the course of the publication. Because of this, serials records tend to be a bit sparse — they don't typically include information about pagination and illustrations, for example — but some information is still required, and if that information changes, it needs to be updated.
Even worse is when two serials merge or one serial splits. In that case, you do have to make new records for the new serials, and then you have to add explanatory links from the old records to the new records, and vice versa. (And then sometimes serials split, only to merge again under a different name.) The ALA used to nominate an annual Worst Serial Title Change Award; when you figure that every single library that subscribed to these journals had to go back and change their records for even a minor title change, it's easy to see why librarians get frustrated when publishers pointlessly tweak serial titles.