| Previous | Table of Contents | Next |
The interworking model illustrated in Figure 17-2 on page 17-5 and Figure 17-3 on
page 17-7 maps a View in one object system to a reference in the other system. This
relationship raises questions:
• How do the concepts of object identity and object life cycle in different object models correspond, and to the extent that they differ, how can they be appropriately mapped?
• How is a View in one system bound to an object reference (and its referent object) in the other system?
5. MD5 was chosen as the hash algorithm because of its uniformity of distribution of bits in the hash value and its popularity
for creating unique keys for input text. The algorithm is designed such that on average, half of the output bits change for
each bit change in the input. The original algorithm provides a key with uniform distribution in 128 bits. The modification
used in this specification selects 118 bits. With a uniform distribution, the probability of
drawing k distinct keys (using k distinct inputs) is n!/((n-k)!*nk), where n is the number of
distinct key values (i.e., n=2118). If a million (i.e., k=106) distinct interface repository IDs are passed through the algorithm,
the probability of a collision in any of the keys is less than 1 in 1023.