What is the point of sampling 'time' with inferred clock?

[henrikt-ma]

Why is sample applied to time here?

ModelicaSpecification/chapters/synchronous.tex

Lines 320 to 324 in 016cdbd

	when Clock() then // equations are on the same clock
	x = A * previous(x) + B * u;
	Modelica.Utilities.Streams.print
	("clock ticks at = " + String(sample(time)));
	end when;

It reminds me of how time is also sometimes sampled in the MSL, such as in Modelica.Clocked.RealSignals.TimeBasedSources.Sine:

  simTime = sample(time);

What is the point of sampling time in these ways? If it is pointless, should the specification and the MSL be simplified in order to not give the impression that it is needed?

[HansOlsson]

That specific line isn't that needed, but it is important to have an additional unrelated equation to indicate that:

when Clock() then // equations are on the same clock 
   x = A * previous(x) + B * u; 
   Modelica.Utilities.Streams.print 
     ("clock ticks at = " + String(sample(time))); 
 end when; 

is different from:

when Clock() then
   x = A * previous(x) + B * u; 
end when;

when Clock() then // This equation is on a separate Clock (which is problematic as it can't be inferred etc).
   Modelica.Utilities.Streams.print 
     ("clock ticks at = " + String(sample(time))); 
 end when; 

As for the question itself, time is a built-in continuous-time variable so to use it in a clocked partition it makes sense to sample() it (similarly as any other continuous-time variable) - if you use it.

However, #3090 changed it a bit, so you can use time directly in a clocked partition (and each use of time is seen as a separate "variable") - but the downside is that the partition is then clocked-discretized (similarly as if you used myTime with der(myTime)=1;). To me the clocked-discretized part is advanced stuff.

Therefore, to me its simpler to use sample(time) than having a clocked-discretized partition in these cases; so I wouldn't change it - but possibly add a note in 16.7.2 Connected Components of the Equations and Variables Graph

[henrikt-ma]

OK, some sort of non-normative explanation sounds like a good solution.