The academic institution that I am slightly affiliated with is in a bit of a kerfuffle about which search interface to use for the Online Public Access Catalog (OPAC). There are departmental politics involved as well as the more understandable questions of user comfort, training issues, and convenience. This is also not the first time these kinds of questions have come up. There have been three separate and distinct default search interfaces for this particular OPAC over the past couple of years.
What most strikes me about the conversation, though, is not that there is some spirited discussion about what interface makes the most sense to use, but rather that we do such a bad job of looking at the search experience from anyone else's perspective. I'm not pointing any fingers here and I am just as guilty of it as everybody else, but we tend to assume that everyone uses the OPAC the same way that we do. I find it hard to believe that most users search like those of us who are discussing the interface, especially when the main contributors to the discussion are librarians who have large amounts of familiarity with the tools and know how to get the results we need.
In order to try and temper my own biases I did a few quick searches of the library literature to see if I could find any studies on how library patrons use the OPAC and there is very little out there. I expected dozens of studies and only found one or two and those were generally done at a much more specific or limited level than what I really wanted. The kind of information we need when making search interface decisions is things like what are users generally looking for? Are they coming in with an item in mind? Are they looking for anything on a particular subject? Are they only interested in stuff they can get their hands on today? Are they starting their search in the OPAC, or are they resorting to it after giving up on the web?
Several of those questions illuminate my own search biases so let me put those biases out there for the record right now. I use the OPAC in one primary way. I have an item I have already identified a need for and I want to know if I can get it from a particular library. In the past, as a student, I would also often use it to find items that I could get my hands on today. If I had a paper due soon (most likely tomorrow), or a project I needed to start working on I would do a few searches to see what, if anything, I could find on a topic that was in the library right now. Most of the research I do now is not for papers due tomorrow or things that require immediate answers. For most of it, sooner is better but a little longer with the best answer is best of all. This means that I simply never use a library catalog for primary subject searches. I know that no library carries every possible item and I know that there are databases of items that are far more comprehensive than almost every individual library catalog. I also know that through purchase or interlibrary loan almost every book in the country and world is accessible. So why would I limit myself to one library's catalog? Why wouldn't I start with Amazon or Worldcat or even just one of the biggest single library catalogs like the Library of Congress'? I find it a little incomprehensible that not everyone thinks in this same way. I'm sure other people have different search strategies, but I find it difficult to imagine what they are and that makes it hard to discuss things like interfaces.
Everyone has this problem. It's the same reason that everyone thinks God has the same morals as they do. Almost anytime someone tells you that people "search this way" what they are really saying is "I search this way." Sometimes they may have enough information to say "I search this way and I've seen other people do it too," but that's rarer and a little misleading. People remember the things they want to remember so if you help ten people search the catalog and eight of them have problems with things that don't bother you and two of them have the same problem you did, you're going to remember those two people, not the eight. Another problem with drawing conclusions from personal experience is that the squeaky wheel gets the oil and there is ALWAYS a squeaky wheel. 99% of the population you're serving can be completely happy with, or at least indifferent to, a change you make in the search interface and therefore never express that opinion. All it takes is one or two people, though, who hate the new look or some other aspect of the new interface and they will have a good shot at derailing any change. This is especially true if those one or two people have the same complaints as some of the librarians in charge of making the decision. And there will always be at least one or two people who hate any change. You could develop a chemical that changes the smell of sewage to the smell of roses and somebody would be unhappy about it.
So what's my point? I guess my most important point is to do some user research, whenever possible. Try not to do it with just regular library users, volunteers, and librarians. See if you can get a few friends, neighbors, or random people off the street to try out your interface (or any other web software you're implementing) and tell you what they think. My other point is to try to differentiate between one or two people with minor complaints and the 99% of people who aren't saying anything.
Also, once you've decided to make a change, watch those user trends. If everyone said they loved the new interface, but usage dropped 50% once you put it in place it is probably time to try again. Just because a decision was made that doesn't mean it can't be unmade or changed to be more effective.