Any procedure P can be called "indexed" P[abc](arguments), instead of the usual P(arguments).
In the body of the procedure, the index can be retrieved and used as an extra argument, or can be simply ignored. Most library procedures ignore the index (mainly if it is not expected).
See a simple standard example with showstat(log);
sin[xxx](Pi/2); # ignored
max[abc](22,33,11); # not ignored, for obvious reasons
Error, invalid input: unrecognized max/min[option] abc
It would be convenient to have always warnings when calling an indexed procedure which ignores the index, but it would be inefficient; anyway, this "feature" is as old as Maple I think.