Thermodynamics of Discrete Systems     13
  
1.1.4. Processes and systems 
1.1.4.1. Definition of a process 
Certain authors define a process (a,b) as a pair of states: initial (a) and final (b). 
They are thereby led to distinguish between states which are possible and those 
which are not. Insofar as we limit ourselves to only consider processes which are 
truly observed (physical processes), the discussion of an axiomatization concerning 
impossible processes, being ill-defined, is beyond the scope and objectives of this 
book. 
From a physical perspective, this means that a process can only be defined if the 
initial and boundary conditions are entirely determined during the process. It makes 
no sense to speak of a process which allows us to pass from a state (a) to state (b) 
unless the external conditions which constrain that process are specified. This is no 
longer a mathematical question, but rather a problem related to a determinism which 
amounts to admitting that an initial state (a), well-defined and always subjected to 
the same constraints, will always lead to the same final state (b). It will always be 
possible to relate two given states, under the condition that, on the one hand, 
exchanges with the exterior furnish the necessary physical quantities, and on the 
other, that the internal system processes which redistribute these quantities allow the 
desired distribution of these system quantities to be achieved. For example, it is not 
possible to realize a state consisting of a given mass whose temperature distribution 
comprises a central peak (Figure 1.1) by means of an action at the exterior walls. 
The necessary energy must be directly supplied to the central zone, which must be 
insulated from the adjacent regions. 
By definition, a process is a series of states. This mathematical definition only 
has physical relevance for processes representative of real evolutions. While not 
precluding a choice of states with no link (a rabbit, a carrot, etc.), the obtention of 
physical evolution laws for matter implies a “certain continuity of content” for this 
ensemble of states. The same goes for all practical problems. A process is therefore 
a succession of states which must be uniquely defined. Apart from some exceptions 
(shocks), we will only consider processes comprising a continuous series of states, 
described by variables which must be continuous functions of time. We will 
however allow situations with discontinuities (shocks, shockwaves, deflagration) 
which momentarily violate this continuity condition. 
1.1.4.2.
 The notion of a system 
The notion of a system is a relatively vague one; it is in fact included in that of a 
process: a system is an entity which we consider during a process. As our 
considerations often take a differential form, the system is the principal part (zero 
order) on which we perform differential balances. The notion of a system is not