Information Technology Reference
In-Depth Information
not counted for the “local” or “simple” rules characterizing maximum source
extensions. Without this special handling, we would not be allowed to add any
of the bonds of the right ring. But since the sulfur atom is still extendable, we
add, in the next step, one of the two ring bonds to it, which results in the code
word
S 0-O1 1-C2 2-C3 0-C4
. As one can immediately see from the source
index 0 in the last bond, this code word is not canonical. Therefore we have to
start the process of rebuilding the code word.
First, the sulfur atom is numbered 0 and this already determines three of the
four parts of the description of the newly added bond from the sulfur to the
carbon atom, namely
0-C?
(the still to be assigned destination index is replaced
by a question mark; alternatively it may be set to the next free node index, which
is 1 in this case, as we did it before). This “incomplete” extension is compared to
the first perfect extension in the volatile part. Since the incomplete description
0-O?
of this extension is greater (as carbon precedes oxygen), the position of the
new edge has been found and the description of this edge is appended. Therefore
we have as a code word prefix
S 0-C1
, which forms the new fixed part of the
code word. The three perfect extensions in the volatile part are renumbered
accordingly (the indices of the destination nodes are increased by one) and their
descriptions are appended, yielding the code word
S 0-C1 0-O2 2-C3 3-C4
(as
before, the gray part is volatile).
The next extension adds the other (ring) bond from the sulfur atom to
a carbon atom: we reorganize from
S 0-C1 0-O2 2-C3 3-C4 0-C5
to
S 0-C1 0-C2 0-O3 3-C4 4-C5
. The sixth extension adds another ring bond,
yielding the code word
S 0-C1 0-C2 0-O3 3-C4 4-C5 1-C6
(before reorgani-
zation). This time, the new edge is not inserted before all perfect extensions,
but after the first, because its source node index is greater than that of the first
perfect extension:
S 0-C1 0-C2 0-O3 1-C4 3-C5 5-C6
. This has two effects:
in the first place, the volatile part now consists of only the last two perfect ex-
tensions (as the insertion of a non-perfect extension edge after it renders the
first perfect extension edge fixed). Secondly, the atom with the maximum source
index (from which on extensions are still allowed) is now the one with index 1,
namely the source atom of the added edge.
The next edge that is added is another ring bond and it is inserted before the
volatile part, since its source index is smaller than the source index of the next
perfect extension bond:
S 0-C1 0-C2 0-O3 1-C4 2-C5 3-C6 6-C7
.
The next edge closes the right ring and it is inserted in the middle of the
volatile part:
S 0-C1 0-C2 0-O3 1-C4 2-C5 3-C6 4-C5 6-C7
(since its source
node index is larger than that of the first perfect extension bond in the volatile
part, but smaller than that of the second perfect extension).
The last three edges, that is, the three bonds of the left ring (3 carbons),
are added in the normal order (after the volatile part, or actually simply by
appending to a fixed code word, since adding the first bond of the left ring renders
the last perfect extension fixed). No code word reorganization is necessary in any
of these steps. The final (canonical) code word is:
S 0-C1 0-C2 0-O3 1-C4 2-C5 3-C6 4-C5 6-C7 7-C8 7-C9 8-C9
.
