From James D. Anderson and José Pérez-Carballo “Information Retrieval Design”:
Data versus information, knowledge; varieties of messages
Information retrieval (IR) databases focus on the retrieval of messages more than simple or raw data. Messages are recorded in documents of many varieties in many media and formats. They make use of visual written language and pictorial images, as well as visual texts based on other representation codes (such as musical, choreographic, chemical, and mathematical notation), aural spoken language and musical performances, and even tactile texts for the visually impaired. These messages reflect the knowledge and understandings of the persons, across generations and cultures, that created them. The lines between “data,” “information,” and “knowledge” are fuzzy. These terms will be discussed in section 1.3 on terminology. Data, information, and knowledge can all be encompassed, represented, or reflected in the messages that humans create. Information retrieval databases are designed to describe and organize such messages so that anyone can find messages they need or desire whenever they want them.
Design of IR databases
The design of databases that will help retrieve messages: Its purpose is to help the designer consider all the relevant factors and to choose the best available options. In most cases, there are no single correct or right answers, only better and worse choices for given purposes and persons.
The IR database designer is confronted with an information retrieval problem and that the designer knows quite a bit about this problem. The components of this problem usually consist of:
a. a large enough set of messages so that it is impossible to easily examine all of them, when certain ones or certain types are desired.
b. a group of actual or potential users or clientele who need access to these messages for purposes of business, life-enhancement, entertainment, or similar impelling reasons.