Cryptography Reference
In-Depth Information
Before:
noun
→
Thelma and Louise
what
(
a
1
)
Harry and Louise
what
(
a
2
)
what
→
went shooting. (
a
3
)
went to the hospital. (
a
4
)
Before expansion:
noun
→
Thelma and Louise
what
(
Harry and Louise went shooting. (
a
1
)
a
2
a
3
a
3
+a
4
Harry and Louise went to the hospital. (
a
2
a
4
a
3
+a
4
)
)
what
→
went shooting. (
a
3
)
went to the hospital. (
a
4
)
Here's the same example reworked for contraction. Before:
noun
→
Thelma and Louise
what
(
Harry and Louise went shooting. (
a
1
)
Harry and Louise went to the hospital. (
a
2
)
a
3
)
what
→
went shooting. (
went to the hospital. (
a
4
)
a
5
)
After contraction:
→
a
1
)
noun
Thelma and Louise
what
(
a
2
+
a
3
)
Harry and Louise
what2
(
→
went to the hospital. (
a
4
)
what
went shooting. (
a
5
)
a
2
a
2
+a
3
what2
→
went shooting. (
)
a
3
a
2
+a
3
went to the hospital. (
)
These rules can be expanded arbitrarily to handle all expansions
and contractions. Weightings like this can significantly affect the way
that bits are converted into phrases using Huffman trees. The trees
work perfectly only if the weights are structured correctly, so it is
highly likely that most trees will produce imperfect approximations
of the weights. As the expansions and contractions change the tree
structure, the weights will significantly alter the patterns produced.
7.3.4 Assessing the Theoretical Security of Mimicry
Determining the strengthof mimic functions based on context-free
grammars is not an easy task. There are two basic approaches and
both of them can leave you with doubt. The first is to analyze the
structure of the systemon a theoretical level and use this to compare
it to other systems. This can indicate that it can often be quite hard
to break through the mimicry in these systems, but it can't prove to
