"... In PAGE 18: ... Table2 : Evaluation Results of Axiom Meister We inspected the inferred specifications by hand to collect the numbers of the last two columns. Most BCL classes are relatively self-contained.... ..."

"... In PAGE 17: ...15 Given the extrapolation involved in Table 2, it is useful to undertake a more detailed analysis of the largest 150 firms (by market value of equity) in order to ascertain the reliability of these estimates and the degree to which they correspond to actual tax benefits.16 Table3 summarizes the evidence provided in Appendix Table 1 on the behavior of these firms and yields more precise, and comparable, estimates of the impact of the proliferation of options on the corporate tax base.17 Using hand-collected data on the exercise of options, $78 billion was realized as proceeds from option exercises for just the top 150 firms in 2000.... In PAGE 17: ... This figure corresponds to a mean ratio of proceeds from option exercises to operating cash flow of 29%. Table3 highlights the growing absolute magnitude of the exercises and the growing ratio of those exercises to measures of corporate profitability during the late 1990s. The detail in Appendix Table 1 demonstrates that the use of options is proliferating through firms of all types and are not limited to high- technology or new economy companies.... ..."

"... In PAGE 17: ...15 Given the extrapolation involved in Table 2, it is useful to undertake a more detailed analysis of the largest 150 firms (by market value of equity) in order to ascertain the reliability of these estimates and the degree to which they correspond to actual tax benefits.16 Table3 summarizes the evidence provided in Appendix Table 1 on the behavior of these firms and yields more precise, and comparable, estimates of the impact of the proliferation of options on the corporate tax base.17 Using hand-collected data on the exercise of options, $78 billion was realized as proceeds from option exercises for just the top 150 firms in 2000.... In PAGE 17: ... This figure corresponds to a mean ratio of proceeds from option exercises to operating cash flow of 29%. Table3 highlights the growing absolute magnitude of the exercises and the growing ratio of those exercises to measures of corporate profitability during the late 1990s. The detail in Appendix Table 1 demonstrates that the use of options is proliferating through firms of all types and are not limited to high- technology or new economy companies.... ..."

"... In PAGE 3: ... We apply the cords statistics-recommendation algorithm that ranks the discovered correlations and soft fds based on their benefit to the op- timizer. For example, Table1 gives the rank- ing of the discovered correlations in a synthetic database of car-accident data. Although cords discovers a correlation between the two columns ACCIDENTS.... In PAGE 3: ... cords does not recommend col- lection of statistics for this pair of columns, since the presence of such statistics would not change any optimizer cardinality estimates. On the other hand, collecting statistics on the first pair of col- umns in Table1 adjusts optimizer estimates by a factor of roughly 3070. 4.... ..."

"... In PAGE 6: ... We collected traces of user input, process information, and network traf c from the six computers. A summary of the collected traces is shown in Table1 . On one hand, these computers were used for daily work, so the traces are real-world.... In PAGE 6: ... On the other hand, our experimental population is small because it is dif cult to convince users to submit their working environment to ex- perimental software. However, from the summary of the collected traces in Table1 , we see that they have good di- versity with respect to hardware, operating system, and user behavior. For real world experiments, we discuss parameter selection and then analyze the performance of our approach on false positives and false negatives.... ..."

"... In PAGE 6: ... We collected traces of user input, process information, and network traf c from the six computers. A summary of the collected traces is shown in Table1 . On one hand, these computers were used for daily work, so the traces are real-world.... In PAGE 6: ... On the other hand, our experimental population is small because it is dif cult to convince users to submit their working environment to ex- perimental software. However, from the summary of the collected traces in Table1 , we see that they have good di- versity with respect to hardware, operating system, and user behavior. For real world experiments, we discuss parameter selection and then analyze the performance of our approach on false positives and false negatives.... ..."

"... In PAGE 5: ... DIFFERENCES BETWEEN PHYSICAL LIBRARIES AND THE INTERNET There are several differences between the traditional library and the Internet. Of the differences outlined in Table1... In PAGE 6: ... ***In certain conditions, pages that have moved have forwarding links, but these links are generally short-lived. ****Links that no longer function Table1 . Differences between Traditional Libraries and the Internet QUALITY CONTROL AND THE INTERNET Libraries have long been a source of educational and entertainment materials, although the collection of a library is generally reflective of the audience to which it is targeted.... In PAGE 7: ... Search engines do not have all Internet sites available, and often report pages that fit the search criteria but are no longer available or are out of date. OTHER DIFFERENCES BETWEEN PHYSICAL LIBRARIES AND THE INTERNET The remaining items in Table1 are also important to consider when defining an Internet research strategy. The lack of a librarian means that the researcher has no one available to help locate items, or to explain various holdings of the Internet.... In PAGE 31: ... These show number of games per table or per croupier for a given time period. RESULTS OF DATA COLLECTED BY HAND The analysis of the data collected by hand is shown in Table1 . As Table 1 indicates, the data passed the chi-square test and runs test.... In PAGE 72: ... psge 69 Proceedings of the American Society of Business and Behavioral Sciences, Volume 7, Number 5 Las Vegas, 2000 Results Research Question One Various questions were asked concerning the purposes of using the Internet for purchasing products or services. Table1 shows the percentage of subjects using the Internet for purchasing, checking product information, or company information. Table 1 The Frequency Information for the Use of Internet... In PAGE 72: ... In the chi-square test, the observed frequency and expected frequency are compared. In Table1 , the chi- square is 7.... ..."

"... In PAGE 5: ... On the other hand we have used the examples collected from the web, as explained in Section 2. Table1 shows, among other information, the number of examples available for each of the words considered. A range of clustering parameters was tried on these sets of examples, as explained on Section 4.... In PAGE 6: ...lustering solutions, e.g. when all the word senses form a single cluster, or all clusters are formed by a single word sense. Table1 shows the 20 nouns that had non-trivial clusters and could therefore be used for evaluation. The quality of a clustering solution was measured using two different metrics that looked at the gold-standard labels of the word senses assigned to each cluster (Zhao amp; Karypis 2001).... ..."

"... In PAGE 2: ... Table1 : Mixer Corpus collection statistics at the conclusion of its first phase Each speaker is asked in each call to specify the phone transmission type (cellular, cordless, or regular landline) and the handset type (speakerphone, headset, ear-bud, or hand-held). This self-reported information could later prove valuable in sorting out the effects of these factors on recognition performance.... In PAGE 3: ... Tables 2, 3, and 4 provide some statistics (which are to be regarded as provisional) on the numbers of speakers and conversation sides from the Mixer collection used in this evaluation. Note that these figures are for conversation sides, while those in Table1 are for whole conversations. Table 2 shows that sizable numbers of speakers and conversations will be included for each language, with a total of 304 different speakers being used.... ..."

"... In PAGE 2: ... Table1 : Mixer Corpus collection statistics at the conclusion of its first phase Each speaker is asked in each call to specify the phone transmission type (cellular, cordless, or regular landline) and the handset type (speakerphone, headset, ear-bud, or hand-held). This self-reported information could later prove valuable in sorting out the effects of these factors on recognition performance.... In PAGE 3: ... Tables 2, 3, and 4 provide some statistics (which are to be regarded as provisional) on the numbers of speakers and conversation sides from the Mixer collection used in this evaluation. Note that these figures are for conversation sides, while those in Table1 are for whole conversations. Table 2 shows that sizable numbers of speakers and conversations will be included for each language, with a total of 304 different speakers being used.... ..."