• English only

# Differences

This shows you the differences between two versions of the page.

sav07_lecture_7 [2009/03/24 23:18]
vkuncak
sav07_lecture_7 [2009/04/08 01:26] (current)
vkuncak
Line 120: Line 120:

The fact that we only used conjunctions is our analysis makes it **path-insensitive**. Intuitively,​ since we're only taking conjunctions,​ our normalization is too conservative and we will not be able to retain properties associated to certain paths. **Path-sensitivity** is often used as a measure of the precision of an analysis. The fact that we only used conjunctions is our analysis makes it **path-insensitive**. Intuitively,​ since we're only taking conjunctions,​ our normalization is too conservative and we will not be able to retain properties associated to certain paths. **Path-sensitivity** is often used as a measure of the precision of an analysis.
+

Line 133: Line 134:
$A^+ = \{ \bigvee (P_{i1} \wedge ... \wedge P_{in}) | P_{ij} \in \{Q_1, ..., Q_n\}) \}$ $A^+ = \{ \bigvee (P_{i1} \wedge ... \wedge P_{in}) | P_{ij} \in \{Q_1, ..., Q_n\}) \}$

-The use of disjunctions allows us to write predicates about attributes for some particular paths. For this reason, this is called **regional** (or **dependent**) **attribute analysis**.+The use of disjunctions allows us to write predicates about attributes for some particular paths. For this reason, this is called **relational** (or **dependent**) **attribute analysis**.

If we go back to our previous example, we could end up with a predicate at the final point which would look (at least partially) somehow like: If we go back to our previous example, we could end up with a predicate at the final point which would look (at least partially) somehow like: