This description from the point of view of authorship places traditional publication at the static end and electronic publication at the fluid end of a spectrum. The reverse picture emerges when analysis turns to receivership, the objective study of rates of access to documents. This study can be conducted without any reference to the actual content of the information resource, though one may wish to eventually connect the two.
Traditional journal articles have a definite life history. There are published, perhaps cited for a while, and then essentially forgotten qua document. By tracking the record of citations to articles, one can obtain a reasonably precise picture of the "biology" of scientific documents [5]. Electronic documents in the first category described above may be expected to obey a life-cycle similar to that of a journal publication. For most of these electronic articles, one can expect the number of "citations" (ftp downloads and http requests) to rise rapidly following the announcement of the article and to fall to zero on a time course determined by any number of variables including the intrinsic value of the article, its timeliness, uniqueness, and so on. Authors have some margin to intervene, by continuously citing their own work [4], for instance. but the process is essentially autonomous. To draw a physical analogy, such documents behavior like a gas: they are released, diffuse, and are finally absorbed into the general cultural substrate.
A objective database, such as a list of references covering a given field, or a collection of DNA sequences, should have a quite different citation time course. After an initial rise in citations after its announcements, the citation rate should be mainly determined by the rate of activity in the field the database refers to, assuming that it is properly maintained. To pursue the physical analogy, such an information resource behaves like a solid: it is reliable and structured and can be expected to serve well as a foundation for any subsequent information-construction activities. An information resource in this category could be distinguished from a resource in the "article" category by the statistical character of its access history, in a purely automatic fashion.
It is inevitable, then, that there exist, or will soon exist, information resources in a third category, neither solid or gaseous, neither fleeting nor immutable. How are these to be identified? How might they be constructed? The case studies presented here aim to provide a framework for answering these questions.