Well, the short answer, for IA in IBM at least is this:
Information Architecture is the practice of organizing, structuring, and classifying information systems, including navigation, linking, and language systems, such that the information is easy to use, easy to find, and easy to understand.
Information architects are often compared to librarians, who catalogue and categorise large amounts of information, usually in book form, in a way that it is easy for people to find. This is usually in a fairly constrained form, as the information is contained within a book, or a journal, or a CD, or DVD etc, and a standard categorisation system is used to determine a location for the information to make it retrievable. Information architects on the other hand often have to deal with information that is less constrained, in multiple formats, and that often has relationship to other material that is ‘out there somewhere’, with no standard categorisation system in which to place the information for ease of retrieval. There’s a tough challenge there! One of the advantages that information architects do have though is being able to influence the development and structure of the information that they need to structure and classify.
Another difference is between the end uses of the information and the goals of the individuals looking for that information. Central to Information Architecture is looking at the roles and goals of the users and ensuring that information strategy is a part of the total user experience. When developing software, the documentation and other associated information is sometimes seen as a secondary part of the product. As perhaps less critical than the software itself. Whereas the information is in fact an important part of the total user experience. The total user experience covers everything from finding about some software to how to upgrade to the next version. And information can be involved in every step. For example: marketing material – why do I want to buy?, instructions on how to order the software, how to get it, how to install it, learning how to use it, how to configure it, how to maintain it, how to expand it, how to make it perform better, how to customise it, how to work out what’s gone wrong, how to service it, how to upgrade it and so on.. At each of these stages the users or user roles needing the information are likely to be different, and have different requirements. Individuals also have preferences for how they retrieve and use the information. Users can also find the information from a variety of angles and locations. Are they reading it from a printed PDF, using an on-line information center, from a link on a wizard, through a search using Google? These have to be considerations in presenting, structuring and classifying information. You have to take a big picture view and understand both the requirements for the users and for the software that is developed for them. You have to understand what information is available, or what information needs to be available, and what the relationship is between the different parts. Most importantly of all, what are the users trying to achieve, what are their goals? How can you best provide the information they need and make it easy for them to find?