The first objective of this paper is to show that some basic concepts used in global navigation satellite systems (GNSS) are similar to those introduced in Fourier synthesis for handling some phase calibration problems. In experimental astronomy, the latter are at the heart of what is called 'phase closure imaging.' In both cases, the analysis of the related structures appeals to the algebraic graph theory and the algebraic number theory. For example, the estimable functions of carrier-phase ambiguities, which were introduced in GNSS to correct some rank defects of the undifferenced equations, prove to be 'closure-phase ambiguities:' the so-called 'closure-delay' (CD) ambiguities. The notion of closure delay thus generalizes that of double difference (DD). The other estimable functional variables involved in the phase and code undifferenced equations are the receiver and satellite pseudo-clock biases. A related application, which corresponds to the second objective of this paper, concerns the definition of the clock information to be broadcasted to the network users for their precise point positioning (PPP). It is shown that this positioning can be achieved by simply having access to the satellite pseudo-clock biases. For simplicity, the study is restricted to relatively small networks. Concerning the phase for example, these biases then include five components: a frequency-dependent satellite-clock error, a tropospheric satellite delay, an ionospheric satellite delay, an initial satellite phase, and an integer satellite ambiguity. The form of the PPP equations to be solved by the network user is then similar to that of the traditional PPP equations. As soon as the CD ambiguities are fixed and validated, an operation which can be performed in real time via appropriate decorrelation techniques, estimates of these float biases can be immediately obtained. No other ambiguity is to be fixed. The satellite pseudo-clock biases can thus be obtained in real time. This is not the case for the satellite-clock biases. The third objective of this paper is to make the link between the CD approach and the GNSS methods based on the notion of double difference. In particular, it is shown that the information provided by a maximum set of independent DDs may not reach that of a complete set of CDs. The corresponding defect is analyzed. One of the main results of the corresponding analysis concerns the DD-CD relationship. In particular, it is shown that the DD ambiguities, once they have been fixed and validated, can be used as input data in the 'undifferenced CD equations.' The corresponding algebraic operations are described. The satellite pseudo-clock biases can therefore be also obtained via particular methods in which the notion of double differencing is involved.