The development of bibliometrics began in the first decades of the 20th century. At that time, scientific research activities had begun to establish themselves. Development of science itself was wanted to be analysed and modelled. Because research findings present themselves above all as publications, it was considered that analysing publications could provide information also about the development of science. The interest towards publications was purely academic at this point.
Bibliometrics lived a low profile for a long time within researchers' chambers, but in the 2000s, it emerged again. Academic ranking systems, scientific recruitments, tenure track arrangements, person evaluations, funding resolutions and such were wanted to be based on "objective" information. Then the possibilities of bibliometrics were noted. Sources of bibliometric information had developed to be easily used. Indicators could be produced easily and swiftly in large numbers.
Nevertheless, indicators are only numbers, and they need to be interpreted in some way in order to be understood. This often has difficulties. What is "small" or "large" in some field of science, and who defines the size or greatness? Who interprets these numbers in the first place and for what purpose? An individual scientist applying for funding? The faculty administration for outcome evaluation meetings? The university management figuring out the worldwide ranking of their institute?
Using bibliometrics to sort out the mutual "rankings" of different fields of science is a perilous choice. Differences can be explained, among others, with different publishing and citing conventions of different fields. The main rule is that a bibliometric analysis is valid only within one field of science. The "Leiden Manifesto for Research Metrics" consists of ten principles that should be followed when evaluating research using bibliometric methods.