JLS 8.3.1.4: contradiction in description of volatile fields

Name: laC46010 Date: 12/21/99


There is a contradiction between two sentences of JLS 8.3.1.4 (p. 148):
    "Another approach would be to declare i and j
    to be volatile:
    class Test {
        static volatile int i = 0, j = 0;
        static void one() { i++; j++; }
        static void two() {
          System.out.println("i=" + i + " j=" + j);
        }
    }
    This allows method one and method two to be executed concurrently,
    but guarantees that accesses to the shared values for i and j occur
    exactly as many times, and in exactly the same order, as they
    appear to occur during execution of the program text by each
    thread. Therefore, method two never observes a value for j greater
    than that for i, because each update to i must be reflected in the
    shared value for i before the update to j occurs. It is possible,
    however, that any given invocation of method two might observe a
    value for j that is much greater than the value observed for i,
    because method one might be executed many times between the moment
    when method two fetches the value of i and the moment when method
    two fetches the value of j."

There is a contradiction between sentences "method two never observes a
value for j greater than that for i" and "that any given invocation of
method two might observe a value for Ï that is much greater than the
value observed for i". It seems to me there are two different
meaninigs of the verb "observe". In the first sentence it might be read
as "can see in main memory", while occurence in the second sentence
might be understood as "have in working copy", or even "print". Such
sense of the sentences quite coincides with the rules for volatile
variables in section 17.7 (p. 407)

======================================================================