### Table 7. Distribution of captured allocation sites within loops under the closed world assumption

Cited by 1

### Table 9. Comparison of captured object allocation sites that could be stack allocated given static stack frames versus dynamic stack frames in both the closed and open world assumptions

"... In PAGE 15: ... In either case, these simpler solutions do not produce dynamic allocation results comparable to published work since they stack-allocate more objects. Table9 illustrates the difference between how many objects could be allo- cated on a static frame size stack versus a dynamic frame size stack. The static frame size numbers are determined by finding all allocation sites that are not contained within a loop and are not array allocations.... ..."

Cited by 1

### Table 1: Experiments for learning operational concepts.

1997

"... In PAGE 16: ... GRDT describes the results when additionally applying the closed world assumption using GRDT. Table1 shows the coverage and accuracy of the di erent learning results. The results strongly support our new representation language coverage and correctness of both learner using the new representation are higher than in the MLJ-experiments.... ..."

### Table 1. The relation RABC stating prior knowledge about the possible combina- tions of attribute values

"... In PAGE 3: ... Thus the underlying domain of our example is the Cartesian product dom(A) dom(B) dom(C) or, as we will write as an abbreviation, the three-dimensional space fA; B; Cg. Table1 states prior knowledge about the possible combinations of at- tribute values in the form of a relation RABC: only the value combinations contained in RABC are possible. (This relation is to be interpreted under the closed world assumption, i.... ..."

### Table 1: Experiments for learning operational concepts.

1997

"... In PAGE 16: ... GRDT describes the results when additionally applying the closed world assumption using GRDT. Table1 shows the coverage and accuracy of the di#0Berent learning results. The results strongly support our new representation language coverage and correctness of both learner using the new representation are higher than in the MLJ-experiments.... ..."

### Table 1. Closed-world results

2000

Cited by 21

### Table 1. Closed-world results

2000

Cited by 21

### Table 1. Knowledge augmentation: open-world assumption

12

"... In PAGE 13: ...nknown. The truth values are shown above the context names. The probability of a truth value of a proposition is defined as the sum over the probabilities of the possible worlds, where the proposition has the truth value (see [24]). For example, consider nine possible worlds DBBDBN BM BM BM BN DBBL in Table1 , one world for each part depicted in Figure 4. The truth value assignment follows the open-world assumption; i.... In PAGE 14: ... In each world, d1 contributes unknown to the augmented truth value. Table1 shows the truth values of the supercontexts (we omit the truth value unknown associated with d1). By considering truth values as the sets true= CUD8CV, false= CUCUCV, inconsistent= CUD8BN CUCV and unknown= CUCV, the union of the truth values defines the truth value within the augmented context d1(s1,s2).... ..."

### Table 1 Federated vs. mediated approach federated approach mediated approach

"... In PAGE 4: ... There apparently is no useful assumption of a closed world in mediated systems due to their dynamics. Table1 summarizes some of those key differences between federated and medi- ated approaches, which also affect security considerations. These differences as well as the similarities of the two approaches are visualized by Fig.... ..."

### Table 6: Number of Redundant Loads Removed Statically

"... In PAGE 6: ...4.2 Impact of TBAA on RLE Table6 gives the number of access paths that RLE removes statically in each of our benchmark programs for each variant of TBAA: TypeDecl, FieldTypeDecl, and SMFieldTypeRefs, By comparing Table 6 and Table 5, we see that the differ- ences between the number of local alias pairs is the strongest indicator of optimization opportunities for RLE. In partic- ular, the big differences between the number of alias pairs for TypeDecl and FieldTypeDecl result in an increase in the number of redundant loads found by RLE.... In PAGE 6: ...4.2 Impact of TBAA on RLE Table 6 gives the number of access paths that RLE removes statically in each of our benchmark programs for each variant of TBAA: TypeDecl, FieldTypeDecl, and SMFieldTypeRefs, By comparing Table6 and Table 5, we see that the differ- ences between the number of local alias pairs is the strongest indicator of optimization opportunities for RLE. In partic- ular, the big differences between the number of alias pairs for TypeDecl and FieldTypeDecl result in an increase in the number of redundant loads found by RLE.... In PAGE 7: ... Since RLE is just one of many optimizations that benefits from alias analysis, the full impact of alias analysis on execution time should be higher. Also, contrary to what the data in Table 5 and Table6 suggest, the three variants of TBAA have roughly the same performance as far as RLE is concerned. These results make two important points.... In PAGE 9: ... Notice that in our experiments, the open-world assumption has an insignificant impact on the effectiveness of TBAA with respect to RLE. This result however reflects the results in Table6 , since SM- FieldTypeRefs, which is most affected by the open world as- sumption, does not enable any additional opportunities for RLE over FieldTypeDecl. With respect to the static metrics, we found that they were the same for the open-world and closed-world assumptions with one difference: M3CG had about 80 more alias pairs (interprocedurally) with the open- world assumption than with the closed world assumption.... ..."