### Table I. Statistics for Non-Layered Implementation of Class FoMission Pure Java JavaSM

### Table 1. Statistics for Non-Layered Implementation of Class FoMission

"... In PAGE 19: ... Other subclasses of MissionImpl encapsulate addi- tions to that are specific to other OPFACs. The Pure Java columns of Table1 present complexity statistics of the FoMission and MissionImpl classes. Note that our sta- tistics for subclasses, by definition, must be no less than those of their superclasses (because the complexity of superclasses is inherited).... In PAGE 19: ...8 Now consider the effects of using JavaSM. The JavaSM columns of Table1 show corresponding statistics, where state exit and enter declarations and edge declarations are treated as (equivalent in complexity to) method declarations. We call such declara- tions method-equivalents.... In PAGE 19: ... We call such declara- tions method-equivalents. Comparing the corresponding columns in Table1 , we see that coding in JavaSM reduces software complexity by a factor of 2. That is, the num- ber of method-equivalents is reduced by a factor of 2 (from 119 to 56), the number of lines of code is reduced by a factor of 3 (from 490 to 143), and the number of symbols is reduced by a factor of 2 (from 3737 to 1615).... In PAGE 20: ... The rows where MissionImpl and FoMission data are listed in bold represent classes that are the terminals of their respective refinement chains. These rows corre- spond to the rows in Table1 . The Isolated Complexity columns of Table 2 show complexity statistics for individual classes of Table 2 (i.... ..."

### Table 1: Traffic in each network segment generated by each scenario. The overhead of the non-layered scenario is highlighted at the bottom, negative values are shown in parenthesis. Traffic is measured in kilobits per frame.

### Table 3. Times required for layout drawing and interconnect tree extraction in SysRel.

### Table 1: Drawing Performance. We show two performance measures for our drawing and culling algorithm: the time to draw the scene with vs. without culling, and the number of nodes drawn (with culling). Our new method achieves a near-constant drawing time for any size tree. Note that for the largest tree, we are not able to enqueue all nodes in the tree as the size of the vector grows too large for the applica- tion to t in main memory.

"... In PAGE 5: ... Both drawing and culling are now based entirely on tree topology, and do not depend on any properties of Grid- Cells. In addition, there is now little bene t to drawing items in order of importance , because our drawing algo- rithm requires a nearly constant amount of work after the dataset size surpass a particular point ( Table1 ). We elimi- nate the extreme performance penalty of updating a sorted queue of items to be drawn, instead using a simple FIFO.... ..."

### Table 1.8: The time complexity of some fundamental graph drawing problems: trees. We denote with k a xed constant such that k 1. Class of Graphs Problem Time Complexity Source tree draw as the Euclidean minimum spanning tree of a set of points in the plane

1997

Cited by 14

### TABLE 44.3.3 The time complexity of some fundamental graph drawing problems: trees. k is a xed constant k 1. CLASS OF GRAPHS PROBLEM TIME COMPLEXITY tree draw as the Euclidean minimum span- ning tree of a set of points in the plane NP-hard

### Table 12. Interaction of lexical, syntactic and semantic information in KPs.

2004

"... In PAGE 10: ... Semantic information in KPs. KP Semantic Link Semantic category Expected variations is a amp;home for Synonymy Home house, living-place, roof are ^colour Hypernym Colour brown, blue, green is a ^time when Hypernym Time period, moment is an amp;amount of Synonymy Amount Quantity Although each kind of token has a value when used alone Table12 shows some possible interactions, and this is that type of search that would include lexical, syntac- tic and semantic information, providing a quite flexible way of getting access to ... ..."

### Table 6. Drawing Time Statistics

1998

Cited by 4

### Table 1: Number of multiplications for a (univariate) single draw

"... In PAGE 13: ... After the Kalman lter, Algorithm 1 applies standard disturbance smoothing whereas method JS applies either equation (3) or (4) or (5) in de Jong and Shephard (1995) which is similar to backwards disturbance smoothing but is computationally more involved. Table1 presents the numbers of multiplications required for a single draw of univariate (p = 1) state space models with di erent state vector dimensions. It is assumed that the elements of Zt, Tt and Rt are either zero or one and variance matrices Ht and Qt are diagonal.... In PAGE 13: ... The computational gains for the modi ed algorithm of x2 3 are virtually the same since the main di erence from Algorithm 1 is that the Kalman lter equation for at+1 in (3) is replaced by the equation for xt+1 in (7) and the resulting di erence is negligible. [ Table1 about here] 3 2 Linear Gaussian illustration In this illustration we follow Fruhwirth-Schnatter (1994a) in considering data from the study of Harvey and Durbin (1986) on the e ect of the seat belt law on road accidents in Great Britain using a Bayesian analysis based on a structural time series model. A graph of the log of monthly number of car drivers killed or seriously injured shows a seasonal pattern due primarily to weather conditions and festive celebrations.... ..."