and Kumar [1982].
高三年级英语第二次调研考试(含答案)
试卷类型：A2020年高三年级第二次调研考试英语2020.6 试卷共8页，卷面满分120分，折算成135分计入总分。
考试用时120分钟。
注意事项：1. 答题前，先将自己的姓名、准考证号填写在答题卡上，并将准考证号条形码粘贴在答题卡上的指定位置。
用2B铅笔将答题卡上试卷类型A后的方框涂黑。
2. 选择题的作答：每小题选出答案后，用2B铅笔把答题卡上对应题目的答案标号涂黑。
写在试题卷、草稿纸和答题卡上的非答题区域均无效。
3. 非选择题的作答：用签字笔直接答在答题卡上对应的答题区域内。
写在试题卷、草稿纸和答题卡上的非答题区域均无效。
4. 考试结束后，请将本试题卷和答题卡一并上交。
第二部分阅读理解（共两节，满分40分）第一节（共15小题；每小题2分，满分30分）阅读下列短文，从每题所给的A、B、C和D四个选项中，选出最佳选项，并在答题卡上将该项涂黑。
ALife under lockdown is strange and lonely but people are finding ways to come together and support those in need by volunteering. Here are just a few more ways you can take part in.NHS RespondersAfter asking retired medical professionals to return to work, the government has also put out a call for 250,000 volunteers to support the NHS （National Health System）. You can apply to be a community response volunteer, collecting and delivering supplies for those in isolation （隔离）；a patient transport volunteer, helping patients who are dismissed from hospital get home and settled; and an NHS transport volunteer, to assist with delivery of medical supplies.GAILY BakeryAs well as baking breads and pastries for the general public, GAILs is supplying front-line NHS teams with fresh food, and they need more help to get it where it needs to be. If you want to lend a hand, drop them an email.Re-engageRe-engage is a voluntary group that helps reduce loneliness among old people and given that so many people are isolated because of the virus outbreak, they need more volunteers to become call companions. If you want to join them in the telephone befriending service, contact them at 1811 2256.A Plate For LondonThis online platform helps match volunteers with Londoners in need. Now they are focusing on doorstepdeliveries of food parcels and hot meals to children who rely on free school meals and those who have lost their jobs as a result of the widespread disease.21. What is the duty of a patient transport volunteer?A. Picking up patients.B. Sending patients home.C. Moving patients to clinics.D. Delivering supplies for patients.22. Which volunteer group especially cares for the senior?A. NHS Responders.B. GAIL's Bakery.C. Re-engage.D. A Plate For London.23. Who may be offered food by the online platform?A. The aged.B. The volunteers.C. The patients.D. The unemployed.BSchool is still out for the summer, but at Eastern Senior High School, students are hard at work. In a garden bursting with vegetables and herbs, nearly a dozen teenagers are harvesting them for the weekend's farmers market. They work Tuesday through Saturday with City Blossoms, a non-profit that brings community gardens to schools and other places where kids gather in urban areas.Roshawn Little, going into her junior year at Eastern, believes that working in the garden has taught her to try all sorts of new things — like eating different kinds of vegetables more often. And she has taken those healthy behaviors home with her and her family now buy more fruits and vegetables. “We mainly live around liquor （酒）stores and snack stores. There aren't that many grocery stores. They're way out, and you have to drive so far" ——a common problem in low-income urban areas .City Blossoms is one of many groups across the country teaming up with local communities to establish school gardens, like the one at Eastern. These gardens, advocates say, are really outdoor classrooms where kids learn valuable lessons — not just about nutrition, but also about science and math, even business skills. For example, the gardens can be used for math lessons —like calculating the area of a plant bed —or learning the science of how plants grow.On a recent weekend at the Aya farmers market, the kids work at a table decorated with handmade signs that read "onions" and "garlic", inviting people to try their herbed salt with bread. Working at the market helps them practice public speaking skills and business skills.Nadine Joyner of a nutrition education company has a food table next to the kids' at the market and often buys produce from them. She is constantly impressed by the kids' knowledge of what they're selling — they know how to grow it, how to prepare it, and how to cook it.24. What do students at Eastern do with the harvested vegetables?A. Sell them on market.B. Take them to school.C. Give them to fanners.D. Share them at home.25. What can we infer about the area Little lives in?A. It is inconvenient in traffic.B. It is crowded with grocery stores.C. It has poor access to fresh foods.D. It suffers from income inequality.26. Why does City Blossoms establish school gardens?A. To provide nutritious food.B. To improve classroom conditionsC. To cooperate with local community.D. To create outdoor learning chances.27. Which may be the best title for this text?A. Green Vegetables, Healthy GardenersB. School Garden: a Promise Land in SummerC. Young Gardeners, Knowledgeable MindsD. City Blossoms: an Exciting Garden ProjectCThe terrible El Nino strikes the globe every 2 to 7 years. As warm waters in the tropical （热带的）Pacific Ocean switch eastward and trade winds weaken, the weather pattern stretches through the atmosphere, causing drought in southern Africa, wildfires in South America, and flooding on North America's Pacific coast. Climate scientists have struggled to predict El Nino events more than 1 year in advance, but artificial intelligence （Al） can now extend the latest forecasts to 18 months, according to a new study.The new research uses a type of Al called a convolutional neural network （卷积神经网络）， which is adept in recognizing images. Researchers feed the neural network a large number of training images and the Al becomes skilled at identifying fundamental patterns of those images. For example, the neural network can be trained to recognize cats in photos by identifying characteristics shared by all cats, such as hairs and four legs. Part of the problem with earlier researches on El Nino forecasts is that they rely on a relatively small set of historical statistics for factors such as ocean temperature. To get around this shortage, the scientists fed the program re-creations of historic ocean conditions produced by a set of reliable climate models, ones frequently used for studies of climate change, says the study's lead author, Yoo-Geun Ham, a climate scientist in South Korea. As a result, the scientists could show the computer system not just one set of actual historic data, lasting from 1871 to 1973, but several thousand reproductions of that same data by the climate models.But it's not clear how much real-world benefit will come from pushing forecasts beyond I year, cautions Stephen Zebiak, a climate scientist at Columbia University. ‘The kind of lead time that is actionable is probably less than a year because decision-makers are unlikely to take action further inadvance，” he says.28. What is the advantage of the latest Al forecasts?A. Faster response。
Herrington, Susan. 2010. The nature of McHarg’s science
L a n d s c a p e J o u r n a l 29:1–10I S S N 0277-2426© 2010 b y t h e B o a r d o f R e g e n t s o f t h e U n i v e r s i t y o f W i s c o n s i n S y s t e mecological message to nonprofessionals, pioneering the use of television media with his show The House We Live In, aired by CBS from 1960 to 1961 (Figure 2).3 McHarg made major contributions to landscape architecture, though aspects of these contributions are problemati-cal, particularly as they relate to science.McH arg was an inventor of ecological planning, and he became a champion of ecological design. As he explicitly stated in his best- s elling book, Design with Nature (1969), “It was not only an explanation but also a command” (McHarg 2006c; 1997, 122).4 McHarg’s ma-jor advancements in landscape architecture include the conception of a novel relationship among nature, design, and science; the promotion of the map- o verlay method; and the use of scientifi c theories to measure desired outcomes in the planning and design process. As chair of the resurrected Department of Landscape Architecture at the University of Pennsylvania, he infl u-enced generations of landscape architects. He was hon-ored accordingly with numerous awards and medals,including election as a Fellow to the American Societyof Landscape Architects (1972) and receipt of the Amer-ican Society of Landscape Architect’s Medal (1984), the LaGasse Medal (1988), the H arvard Lifetime Achieve-ment Award (1992), and the Japan Prize (2000).Underlying all of these achievements was McHarg’s belief that science was a truth serum that would revealthe verifiable facts of nature to humans. Science pro-vided not only an explanatory model for understand-ing nature but also a prescriptive one. Equipped with the revelatory powers of science, nature would serve as a guide to design and planning. While McHarg consis-tently substantiated his ecological ideas with scientific theories, he possessed no formal training as a scientist and he never claimed to be one. In fact, he only took one science class in 1938 at the West of Scotland Agricultural College (MY 1976, 105; McH arg 1996, 82). Althoughhe once referred to himself as a quasi-p seudo- c rypto-s cientist, many considered him to be a scientist.5 David Orr has suggested that McHarg was “more of a scientist than many he employed. He was a perceptive observer of the wayward ways of men and their tendency towardsABSTRACT Ian McHarg undoubtedly will be remembered as one of the most infl uential landscape architects of the 20th century. His charismatic personality, grand narrative Design with Nature, and unwavering conviction that science would provide meaning and purpose for landscape architects placed him at the center of debates concerning nature, design, and planning. Yet his visions have been criticized as well as praised. Rarely straying from the ideas he synthesized in the 1960s, McHarg consistently contra-dicted himself. He criticized humans for privileging man over all other considerations, but he himself was autocratic, asserting his views as absolute and superior to all. His vision of nature was that of dynamic process, yet he sought to plot and rank natural phenomena on static maps. In promoting outdated ideas about science as a savior for landscape architecture, he used rhetorical devices suggestive of religious discourse. This paper attempts to unravel the complexity and contradictions of McHarg’s views on science. After reviewing the contributions of McHarg, it examines problems with his assertions relating to the ecological superior-ity of English landscape gardens, promotion of the map- o verlay method as a scientifi c process, and the combination of Lawrence Henderson and Charles Darwin’s work for his theory of creative fi tting.KEYWORDS history, science, design, natureSCIENCE AS A TRUTH SERUM FOR LANDSCAPEARCHITECTURE 1As an author, academic, public personality, and practitioner, Ian McHarg (1920–2001) profoundly changed the teaching and practice of landscape ar-chitecture (Figure 1).2 McH arg revealed the damagedcondition of the natural environment and held the electrifying promise that landscape architects were in-strumental to its repair. H e condemned Renaissance, Baroque, and École des Beaux- A rts formalism and championed the use of natural sciences in environmen-tal design. To raise landscape architecture from what he perceived to be the lowly and wanton ways of garden art, not only did he write and teach about the value of science to design but also, with his offi ce Wallace, McHarg, Roberts, and Todd (WMRT), he set out to use science in the design of regional landscapes. As such, he spearheaded multidisciplinary teams including ex-perts from various scientifi c fi elds, and he advanced the map- o verlay method—a key to his ecological model and a precursor of computerized Geographic Infor-mation Systems (GIS). Notably, he sought to bring his The Nature of Ian McHarg’s ScienceSusan Herrington2Landscape Journal 29:1–10and Lawrence Halprin, but he found most practitioners in North America uninspired and mediocre at best. He sought to change this state of affairs by developing a landscape architecture curriculum that was better than the program at his alma mater, Harvard University. He did this by recruiting top- r anked architecture students and developing courses and studios that included a body of knowledge—the natural sciences—missing from his own experience as a student at Harvard. Sci-entists were frequent guests to the school, and in 1960 McH arg hired Nicholas Muhlenberg, a scientist with background in forestry and ecology, as part of the per-manent faculty (McHarg 1996, 172).In 1959 McHarg started the course “Man and En-vironment,” which involved guest lecturers investi-gating “the scientifi c conceptions of matter, life, and man; the views of God, man, and nature in the major philosophies and religions” and the ecological inter-actions of humans and nature (McHarg 1996, 140). In time, theologians declined invitations to the course. McHarg noted: “Scientifi c expositions amplifi ed the un-derstanding of the miraculous in nature. No scriptural description of the supernatural could remotely com-pare to the scientifi c view” (1996, 161). He also inves-tigated the Potomac River Basin with students as part of his design studio. Building upon his experience with large, collaborative projects at Harvard, McHarg incor-porated numerous scientists into the fold of this multi-disciplinary team. The study integrated an assortmentselfi sh destruction of the environment” (2007, 9). Prob-lems nevertheless emerge from McHarg’s conceptions and use of science.Despite all of McHarg’s triumphs, some aspects of his science were inaccurate. Specifi cally, his ideas re-garding the ecological superiority of English landscape gardens, the promotion of the map- o verlay method as an objective process, and the combining of Lawrence H enderson and Charles Darwin’s scientifi c theories were misguided. Landscape architects and students of landscape architecture continue to perpetuate these in-accuracies. Analyses of McHarg’s texts, projects, and lec-tures are useful in teasing out the problematic strands of this great landscape architect’s ecological message. Ultimately, they demonstrate that science, like history, is continually revised and that only by incorporating these changes into our own body of knowledge can we benefi t from its wisdom.SOME MAJOR CONTRIBUTIONS OF MCHARGScience Links Design to NatureAs an assistant professor at the University of Pennsyl-vania in 1954, McH arg was charged with the task of creating a new landscape architecture curriculum. He viewed the profession as plagued by low self- e steem in both the academic community and in society in gen-eral (McHarg 1996, 129). McHarg admired the modern revolutionaries Christopher Tunnard, Thomas Church,Figure 1. Ian McHarg (Architectural Archives of the University ofPennsylvania).Herrington 3cal Symbol” (1980), and an essay for the American So-ciety of Landscape Architects (2006c; 1997). McHarg’s support of science sometimes assumed a religious fer-vor, and he has been described as the Billy Graham of ecology (Hedgpeth 1986, 48). The opening chapters of Design with Nature (McHarg 1969) take readers from the countryside of Scotland, dune development in the Netherlands, natural disasters, pollution, tacky com-mercial strips and dense urban living (that would in-spire a later generation of designers), and homage to the people of Japan (whom he viewed as indivisible from nature) to a tirade against Western Civilization that valued only individuality, economic determinism, and anthropocentric art. At the end of this blistering critique, he asked, “Where else can we turn for an accu-rate model of the world and ourselves but to science?” (McHarg 1969, 29)Historical Gardens and Ecological IntegrityA recurring feature in McHarg’s texts and lectures was the use of representational analogies to validate design-ing with nature. In making comparisons to show simi-larities, his analogies were central to both the discovery and explanation of scientifi c theories. Most notably, Sir Isaac Newton explained his theory of gravitation by drawing a likeness between the way the earth pulls on an apple falling from a tree and the way it attracts theof scientifi c data, including “meteorology, geology, geo-morphology, groundwater and surface hydrology, soils, vegetation, wildlife, limnology, and, where appropriate, physical oceanography and marine biology” (McH arg 1996, 194). This project was “the fi rst of its kind to use the physiographic region and the river basin as the pri-mary organizing context for ecological planning and design—a framework that linked past, present, and an-ticipated future actions and multiple landscape scales from garden to region” (Spirn 2000, 105). For McHarg, this regional planning study was an “expansion of pro-fessional responsibility” (McHarg 1996, 195) that estab-lished his ecological planning method (1996, 197).In 1963, McH arg contended that the ecological and natural sciences offered an important theoretical framework for landscape architects and planners. In “Man and Environment,” which appeared in Leonard Duhl’s The Urban Condition, McH arg asked readers to trade in older, humanist ideas regarding nature for the scientist’s view of the evolution of nature. He pos-ited: “The inheritors of the Judaic- C hristian- H umanist tradition have received their injunction from Genesis, a man- o riented universe” (McH arg 2006f; 1963, 3). McH arg viewed religious doctrine as separating hu-mans from nature, whereas science provided an inte-grative view of humans and nature. In this way, nature provided a wellspring of truth about the world, and sci-ence, a means to reveal this truth. Relating this idea di-rectly to the design and planning process, he surmised, “We have asked Nature to tell Man what it is, in the way of opportunities and of constraints for all prospective land- u ses” (McHarg 2007, 44).In terse prose, McH arg repeatedly condemned Judeo- C hristian traditions and Western culture in gen-eral as the legitimizing force behind our separation and dominion over nature, and he consistently promoted science as the alternative. This argument appears in his papers “Man and Environment” (2006f; 1963) and “Values, Process, and Form” (2006g; 1968), his speech “Man: Planetary Disease” (1971), his book Design with Nature (1969), his lecture “The Garden as a Metaphysi-Figure 2. Ian McHarg interviewing psychologist Erich Fromm on the set of The House We Live In, October 1960 (Architectural Archives of theUniversity of Pennsylvania).4Landscape Journal29:1–10Figure 3. Snowfl akes are examples of nature’s design (From Wilson Bentley, Studies among the Snow Crystals in Monthly Weather Review. 1902. Plate XIX).Herrington5las Poussin, and Salvator Rosa (1969, 73). Nonetheless, McHarg assured readers, in the English landscape gar-den “the ruling principle was that ‘nature is the garden-er’s best designer’—an empirical ecology” (1969, 73).In advocating for nature, science, and design, McHarg also introduced a novel defi nition of nature. He posited that places are “only comprehensible in terms of physical and biological evolution” (1967, 105) and that nature as a process is subject to the forces that produce and control the phenomena of the biophysical world. This defi nition of nature as a process is a lasting con-tribution. And the method he advocated for integrating natural processes into design and planning was equally enduring. This method used extensive landscape in-ventories that were most notably operationalized by the process of mapping.Map Overlays in the Ecological ModelThe map- o verlay method was key to McH arg’s eco-logical model. This process spatially referenced the inventoried data and weighted its relative importance to design decision- m aking as part of the analysis. Origi-nally, the map- o verlay system involved layers of trans-parent fi lm over a base map. Other types of transparent materials, and eventually the computer, replaced the fi lm overlays. Each layer of fi lm was dedicated to a single inventoried factor, such as topography or historic sites, which was rated from a high to low value. The darkest gradations of tones represented areas with the highest value and the lightest tones indicated areas with the least signifi cant value. All of the mapped layers were then superimposed to create a composite map that in McHarg’s words looked something like a “complex X-ray photograph with dark and light tones” (1969, 35).For McH arg, the composite map was where the truth was revealed. Development suitability was rated on the map from highest (lightest color) to lowest (dark-est). According to McHarg, the integration of social and natural information across the site enabled designers to chart future development in ways that closely adhered to nature’s intrinsic progression towards stability. H e augmented the map analyses with technical reports,moon. McH arg frequently used the frozen hexagonal symmetry of snowfl akes (Figure 3) or the elegant util-ity of a bird’s beak to symbolize the inherent beauty of nature’s designs. Moreover, to represent nature’s design at a larger scale, McHarg consistently referred to 18th-c entury English landscape gardens, which he viewed as representing ecological concepts. Humans creating these gardens were designing with nature, while earlier Western gardens were not designed with nature.McHarg frequently berated Renaissance gardens as the penultimate expressions of Judaic- C hristian tradi-tions and Western culture. He found they “clearly show the imprint of humanist thought. A rigid symmetrical pattern is imposed relentlessly upon a reluctant land-scape” (2006f; 1963, 8). In Design with Nature (1969) his critique of historical landscapes took a full chapter, ti-tled “On Values.” McHarg began with American Indians in North America, who he claimed “evolved a most har-monious balance of man and nature” (1969, 67). He then moved on to the imperious Renaissance gardens, where he perceived “the imposition of a simple Euclidean ge-ometry upon the landscape” (1969, 71). Ultimately, the French Baroque gardens designed by Le Notre were “testimony to the divinity of man and his supremacy over a base and subject nature” (1969, 71).6 Despite the fact that the term ecology was not coined until the 19th century, McHarg found that these gardens had “no ecological concept of community or association” (1969, 71).7 For these reasons, McHarg used them as represen-tational analogies of not designing with nature.Unlike those involved in earlier gardening tradi-tions, McHarg (1969) wrote, a handful of 18th- c entury landscape architects believed that “some unity of man and nature was possible and could not only be created but idealized” (1969, 72). For McH arg, English land-scape gardens were designed by adhering to a site’s natural functions, making them analogies of designing with nature. He claimed: “Never has a society accom-plished such a benefi cent transformation of an entire landscape” (1969, 72). He admitted that the designers of these gardens took their cues as to what nature looked like from the Romantic painters Claude Lorraine, Nico-6Landscape Journal 29:1–10intrinsically suitable locations for various types of devel-opment. The answer was s imple—map it.McHarg’s ecological method was both a “diagnosis and prescription” for development (Palmer 2001, Spirn 2000), and he believed it was an objective procedure that could be replicated to produce the same outcomes. Describing the Woodlands project, to Landscape Ar-chitecture readers in 1975, McHarg and Jonathan Sut-ton noted, “H aving accumulated and interpreted the biophysical data describing the region and [an] 18,000 a[cre] site, a method was developed which insured that anyone would reach the same conclusions . . . any engi-neer, architect, landscape architect, developer, and the client himself were bound by the data and the method” (McHarg and Sutton 1975, 78).Later in life, McH arg was enthusiastic about the importation of his mapping method into computerized, geographic, information systems, exclaiming, “More data can be ingested, evaluated, and synthesized faster, and more accurately than ever before” (2006c; 1997, 119). He surmised that fi nally, the “computer will solve the command ‘show me the locations where all most propitious factors are located and most detrimental factors are absent’” (2006c; 1997, 118). To be sure, the computer fulfi lled his quest for a systematic method of landscape design based on scientifi c data, but McHarg’s approach was also substantiated by scientifi c theories concerning evolution and adaptation.Creative FittingAs a means of lending scientifi c integrity to his eco-logical approach, McHarg developed a scientifi c theory called creative fi tting that both explained and validated designing with nature. McHarg’s method was ecologi-cal not only because it used ecological data but also be-cause the outcomes it produced matched the processes of adaptation and evolution. It helped determine where proposed human uses, such as buildings and roads, intrinsically fi t on the land. Since this design method located the fi ttest environment for various land uses, it also fulfi lled the basic principles of adaptation (2006c; 1997, 124–125).suitability matrices, diagrammatic sections, decision trees, and other techniques to reveal the natural pro-cesses of a site. Map overlay, however, is consistently attributed to McHarg’s ecological method.In the early 19th century, epidemiologists used similar mapping techniques to establish the location of contagions in cities (Tufte 1983, 27). Likewise, McHarg’s mapping method shares commonalities with the over-lays employed by the U.S. Defense Department to de-termine target locations for intercontinental missiles during the Cold War (Cloud 2001, 203). In landscape architecture, the map- o verlay method advanced by McHarg was similar to that used by Warren Manning in the early 20th century to record and classify site infor-mation (Zube 1986; Neckar 1989).8 In 1950 Jacqueline Tyrwhitt provided one of the fi rst explicit descriptions of the map- o verlay method in the design process in her book Town and Country Planning Textbook (Steinitz, Parker, and Jordan 1976, 445). Tyrwhitt was a professor at the Graduate School of Design in the Department of City Planning and Landscape Architecture at Harvard University from 1955 to 1969. While McH arg was not the inventor of the map- o verlay method, he certainly championed it as no other individual before him.McH arg introduced his method to landscape ar-chitects in 1965 in his Landscape Architecture article “Plan for the Valleys vs. Spectre of Uncontrolled Growth” with David Wallace. Examining and layering geologi-cal, topographical, economic, and a multitude of other factors, they demonstrated how planned growth could save seven million dollars as compared to uncontrolled growth (McHarg and W allace 1965, 180). The map- o verlay method also features signifi cantly in Design with Nature, where McHarg structured the text so that chapters rife with condemnations of corrupt cities, Western values, rampant pollution, and urban pathologies alternated with chapters detailing the life- s aving solutions of his map- o verlay process. Many of these projects were com-pleted with University of Pennsylvania students and his offi ce, WMRT. Accompanied by God- l ike aerial views of the earth, these polychromatic maps revealed to read-ers how one might design with nature, thereby locatingHerrington 7most fi t, to adapt these and themselves, continuously . . . Systems which are fi t are evolutionary successes; they are maximum success solutions to fi tness” (2006b; 1978, 87). McHarg’s explanations of the Darwin / H enderson- i nspired theory of creative fi tting stressed that evolu-tionary progress, left to nature, moves towards some optimal point of success—that best evolutionary suc-cess is best defi ned by maximum fi t solutions (McHarg 1969, 120; 1996, 245; 2007, 23).This may be why the living and nonliving things, that were in McH arg’s view ‘unfi t,’ drew his criticism. They did not fulfi ll what he deemed the optimizing directionality of adaptation. In defi ning what is unfi t, McHarg (1969, 170) noted:Our language conforms to this notion of the un-fi t as the unhealthy, crippled, deformed, although there may well be excellences that overcome this. Beethoven transcended deafness. So unfi tness would include not only the broken piano, but also the de-faced painting . . . the house in shade or the glaring street, the anarchic city; these are all unfi t.Stability. Another dimension concerning the outcomesof evolution and adaptation is the idea of stability as a benchmark of ecological health. McH arg argued that “stable and healthy forests, marshes, deserts, streams can be defi ned, that succession and retrogression can be identifi ed” (2006a; 1966, 39). H e further asserted: “Complexity, diversity, stability (steady state), with a high number of species and low entropy are indicators of health and systems moving in this direction are evolv-ing” (1967, 107), and again attested that an increase in the number of species correlates with an increase in stability (2006g; 1968, 57). He concluded that ecologi-cal fi tness must meet the evolutionary criteria of com-plexity, diversity, stability and interdependence (2006g; 1968, 60), and “on all counts the complex environment will be more stable”(1969, 120).Adaptation. In McHarg’s words, the theory of creative fi tting “has absolutely no status whatsoever, except in-sofar as all the parts have been derived from excellent scientists” (2007, 21). Indeed, creative fi tting conjoined the scientifi c theories of Charles Darwin’s The Origin of Species (2003; 1859) and the lesser- k nown scientist Law-rence Henderson’s The Fitness of the Environment (1958; 1913). Creativity, for McHarg, was not an act exclusive to human artists but rather a directional process towards higher levels of order, which he thought occurred in the laws of both thermodynamics and evolution—in living and nonliving systems (2007, 22). McHarg defi ned fi t as a blend of two scientifi c propositions: Charles Darwin’s idea that “the surviving organism is fi t for the environ-ment” (2007, 23) and Lawrence Henderson’s theory that “the actual environment, the actual world, constitutes the fi ttest possible abode for life . . . this fi tting then is essential to survival, according to Darwin, and there is always a most fi t environment for every system seek-ing an environment” (2007, 23–24). In 1981, McH arg wrote, “Every organism, system, constitution, is re-quired to fi nd the fi ttest environment, adapt that envi-ronment and itself in order to survive” (McHarg 1981, 93). In 1997, he again referred to his theory of creative fi tting in a defi nition of ecological design:Ecological design follows planning and introduces thesubject of form. There should be an intrinsically suit-able location, processes with appropriate materials, and forms. Design requires an informed designer with a visual imagination, as well as graphic and creative skills. It selects for creative fi tting revealed in intrinsic and expressive form (McHarg 2006c; 1997, 123).He concluded: “And, thanks to Charles Darwin and Law-rence Henderson, we have a theory” (2006c; 1997, 124).Evolution. Creative Fitting also explained how McHarg’secological method produced outcomes matching the trajectory of evolution. He anchored his theory of fi t in the Darwin / H enderson combination, stating, “All sys-tems are required to seek out the environment that is8Landscape Journal 29:1–10ard Weller (2006) has posited that landscape urbanism must conjoin the rigor and conviction characterizing McH arg’s ecological method with the exquisite imag-ery and theoretical sophistication that defi nes Corner’s work. For Weller, “We are aptly reminded that landscape architecture is at best an art of instrumentality, or bet-ter still, an ecological art of instrumentality” (2006, 77). Indeed landscape urbanism shares commonalities with McH arg’s work, particularly its emphasis on graphic and analytic techniques, and its dependency on sys-tems and strategy over form and design.Critics of McHarg have frequently observed his dis-interest in social issues. As early as 1971, Michael Lau-rie cautioned: “By his own admission McH arg barely touches upon social issues beyond the realm of sur-vival” (206, 248). And yet some of the most serious criti-cism leveled at McHarg concerns his disdain for art and his low regard for site design, particularly at a garden scale. In Recovering Landscape, Marc Treib stated:[McHarg] mixed science with evangelism—a sort of ecofundamentalism . . . McHarg’s method insinuated that if the process were correct, the consequent form would be good, almost as if objective study automati-cally gave rise to an appropriate aesthetic. In responseto his strong personality and ideas, landscape archi-tects jumped aboard the ecological train, becoming analysts rather than creators, and the conscious mak-ing of form and space in the landscape subsequently came to a screeching halt (1999, 31).Given that Treib’s essay is the fi rst chapter of Recov-ering Landscape, readers might wonder whether land-scape architecture is recovering from McH arg. Spirnalso noted in 2000:When McH arg calls ecology “not only an explana-tion, but also a command,” he confl ates ecology as a science (a way of describing the world), ecology as a cause (a mandate for moral action), and ecology as an aesthetic (a norm for beauty). It is important to dis-tinguish the insights ecology yields as a description of the world, on the one hand, from how these insightsPROBLEMS AND LIMITATIONS OF McHARG’S CONTRIBUTIONSReaction to McHarg’s WorkDesign with Nature made its debut in 1969, and since then conceptions of nature, design, and science have developed in landscape architecture in myriad ways. Some of these changes have been the direct conse-quence of reinterpreting McHarg’s work. Without doubt he sustained a devout following, particularly among his past students. The landscape architecture offi ce An-dropogon Associates, founded by Carol Franklin, Colin Franklin, Leslie Sauer, and Rolf Sauer (McHarg’s former students), has produced award- w inning landscapes in the name of “designing with nature.” On the teaching front, Ecology and Design: Frameworks for Learning (Johnson and Hill 2001) has revamped the role of eco-logical thinking in design education; some consider it the postscript of Design with Nature (Pittari 2003, 115).Former students have also challenged McH arg’s autocratic views, expanding the scope and intent of his ecological message. Whereas McH arg viewed the city as the antithesis of nature, his former students, Anne Whiston Spirn and Michael H ough, found the urban environment brimming with natural systems worthy of our attention (Spirn 1984; H ough 1995). McH arg’scolleagues also countered some of his positions on thefuture of landscape architecture, bringing a humanistdimension to his ecological method. McH arg soughtto make landscape design and planning a hard sci-ence (Olin 1999, 16). Laurie Olin, in both his practicewith Robert Hanna and his teaching at the University ofPennsylvania, however, combined ecological concernswith artistic expression. When James Corner joined theUniversity of Pennsylvania in 1988, he intentionally ex-ploited the subjective beauty of maps. With Anu Mathur, Corner argued that maps were part of the repertoire of representational strategies used to generate “a catalytic locale of inventive subterfuges for the making of poetic landscapes” (Corner 1992, 275).Even the latest trend in landscape architecture—landscape urbanism—has ties to McHarg’s work. Rich-。
逃跑的女孩
逃跑的女孩Rajesh Kumar Kanoi【期刊名称】《海外英语（中）》【年(卷),期】2003(000)004【摘要】@@ Ru Yun and San Hua were best friends. They had been the best of friends for almost half their lives, at about fifteen years of age. They went to the same school,took the same bus and shared the same interests. They were always together like a pair of Siamese twins1. Nothing could separate them, not even their studies or their parents,combined ire2, and often justified, too. Perhaps, their parents would not have minded their companionship if they had something to show for it. But, the girls were both lagging in their studies, their results were poor and getting worse with time. They showed no interest or inclination3 for studies and that was the one thing that was most expected of them. Had they even made an attempt their parents would have been more kindly disposed towards them. But, term after term,year after year, they kept slipping until they dropped right to the bottom of their class. If there was any competition between them it was to see who was better, bottom up4.【总页数】2页(P32-33)【作者】Rajesh Kumar Kanoi【作者单位】无【正文语种】中文【相关文献】1.逃跑的火车逃跑的星辰 [J], 尤克利;杜凤刚2.在逃跑中寻找爱与被爱的感觉——由《逃跑新娘》引发的联想 [J], 刘江3.身份协商以及身份的多元化和差异性——评《如何约会一个棕女孩、黑女孩、白女孩或混血女孩》 [J], 王峰4.有钱人逃跑比穷人逃跑更可怕 [J], 萧常5.逃跑:回归原点的身份之旅——图森《逃跑》中的中国人和西方人 [J], 汤娜因版权原因，仅展示原文概要，查看原文内容请购买。
study on Zigbee
Impact Factor: 1.852 IJESRTINTERNATIONAL JOURNAL OF ENGINEERING SCIENCES & RESEARCHTECHNOLOGYStudy on ZIGBEE TechnologyAbhishek Kumar*1, Sandeep Gupta2*1,2Department Of ECE, Bharat Institute of Technology, Partapur, Meerut-250003, India999electro.abhi@AbstractZIGBEE is one of the most widely used transceiver standard in wireless sensor networks. Zigbee over IEEE 802.15.4., defines specifications for low rate WPAN(LR-WPAN) to support lower monitoring and controlling devices. Zigbee is developed by Zigbee alliance ,which has hundreds of member companies. Zigbee alliance(software) defines the network, security and application layers. IEEE802.15.4(hardware) defines the physical and media access control layers for LR-WPAN. This paper presents a detailed study of Zigbee wireless standard, IEEE802.15.4 specification, Zigbee device types, the protocol stack architecture and its application.Keywords: Zigbee, IEEE802.15.4. Standard, LR-WPAN.IntroductionWireless Technology is being developed rapidly nowadays. Advancement in micro electromechanical systems brings integration of sensing, signal processing and RF capability on very small devices. All kind of portable applications tend to be able to communicate without the use of any wires. Aim of wireless communication is to gather information or perform certain task in the environment. A typical sensor node contains three C’s, are Collection, Computation and Communication units. Based on the request of sink, gathered information will be transmitted wirelessly. The collection unit has series of sensors. Computation unit contains microcontroller and memory. Finally the communication unit contains transceiver to transmit and receive data; various transceivers (such as RFM TR1000 family, Hardware accelerators, ChipconCC1000 and CC2420 family , Infineon TDA 525x family, IEEE802.15.4/Ember EM2420 RF transceiver, ConexantRDSSS9M) used for this purpose.The reasons [1] for using Zigbee are,•Reliable and self healing• Supports large number of nodes.•Easy to deploy•Very long battery life•Secure•Low cost•Can be used globally• Vibrant industry support with thirty or more vendors supplying products and services •Open Standards protocol with no or negligible licensing fees•Chipsets available from multiple sources•Remotely upgradeable firmware• No new wires•Low power (ability to operate on batteriesmeasured in years)•Low maintenance (meshing, self organizing)•Standards based security [AES128]•Ability to read gas metersAll of the technologies are young – Bluetooth being the oldest with developments started in 1997. ZigBee started its developments in 2001. Different companies developed other technologies within the last three or four years. Zigbee is one of the most widely utilized Wireless Sensor Network standards with low power, low data rate, low cost and short time delay characteristics, simple to develop and deploy and provides robust security and high data reliability. Name of the Zigbee came from zigzagging patterns of honey bees between flowers, represents the communication between nodes in a mesh network [1].ZIGBEE and IEEE 802.15.4ZigBee is developed by ZigBee alliance, which has hundreds of member companies (Ember, Freescale, Chipcon, Invensys, Mitsubishi, CompXs, AMI Semiconductors, ENQ Semi conductors), from semiconductor and software developers to original equipment manufacturers. ZigBee and 802.15.4 are not the same. ZigBee is a standard based network protocol supported solely by the ZigBee alliance that uses the transport services of the IEEE802.15.4 networkhttp: // (C) International Journal of Engineering Sciences & Research Technology[2733-2738]specification. ZigBee alliance is responsible for ZigBee standard and IEEE is for IEEE802.15.4. It is like TCP/IP using IEEE 802.11b network specification [2]. ZigBee alliance (software) defines the network, security and application layers. IEEE802.15.4 (hardware) defines the physical and media access control layers for LR-WPAN in figure1. Power needed for ZigBee is very small. In most cases it uses 1mW (or less power). But still it provides range up to 150 meters in outdoor which is achieved by the technique called direct sequence spread spectrum (DSSS). Also DSSS consumes less power compared to Frequency Hopping Spread Spectrum (FHSS). It works in the 868 MHz (Europe),915 MHz (North America and Australia) and 2.4 GHz(available worldwide) ISM band with up to 20kbps, 40kbps and 250kbps data rate respectively .Because these wave bands are different from the bands of current common wireless networks, Wireless Fidelity (Wi-Fi), Bluetooth, Wireless USB etc. Mutual interferences between them will not occur, therefore, this guarantees our system will not interfere other wireless networks and will not be affected as well.Figure 1: ZigBee adds network, security, and application-services layers to the PHY and MAC layers of the IEEE811.15.4 radio.The IEEE 802.15.4 standard employs 64-bit and16-bit short addresses to support theoretically more than 65,000 nodes per network [7]. ZigBee network can have up to 653356 devices, the distance between ZigBee devices can be up to 50 meters, and each node can relay data to other nodes. This leads capability of making a very big network which covering significant distances.ZIGBEE StandardZigBee device are the combination of application (such as light sensor, lighting control etc), ZigBee logical(coordinator, router, end device), and ZigBee physical device types (Full Function Device and Reduced Function Device)[1].A,ZigBee physical device types: Based on dataprocessing capabilities, two types of physical devices are provided in IEEE 802.15.4: Full Function Devices (FFD)and Reduced Function Devices(RFD). Full Function Devices can perform all available operations within the standard, including routing mechanism, coordination tasks and sensing task. The FFD plays role of coordinator or router or end devices (It can be either FFD or RFD depends on its intended application). A typical FFD in a ZigBee network will be powered from an AC-fed mains supply, as it must always be active and listening to the network . Reduced Function Devices, on the other hand,implements a limited version of the IEEE 802.15.4 protocol. The RFDs do not route packets and must be associated with an FFD. These are end devices such as sensors actuators which only doing limited tasks like recording temperature data, monitoring lighting condition or controlling external devices. The current ZigBee standard requires FFDs to be always on, which in practice means that FFDs must be constantly powered. Battery-powered FFDs have a lifetime on the order of a few days. B. ZigBee logical device types :There are three categories of nodes in a ZigBee system.They are Coordinator, Router and End devices. 1) Coordinator : Forms the root of the network tree and might bridge to other networks. There is exactly one coordinator in each network. It is responsible for initiating the network and selecting the network parameters such as radio frequency channel, unique network identifier and setting other operational parameters. It can also store the information about network, security keys.2) Router: Router acts as intermediate nodes, relaying data from other devices. Router can connect to an already existent network, also able to accept connections from other devices and be some kind of re- transmitters to the network. Network may be extended through the use ofZigBee routers.Figure 2: Zigbee Networkhttp: // (C) International Journal of Engineering Sciences & Research Technology[2733-2738]3) End Devices : End Device can be low-power/ battery-powered devices. They can collect various information from sensors and switches. They have sufficient functionality to talk to their parents (either the coordinator or a router) and cannot relay data from other devices. This reduced functionality allows for the potential to reduce their cost. They support better low power models. These devices do not have to stay awake the whole time, while the devices belonging to the other two categories have to. Each end device can have up to 240 end nodes which are separate applications sharing the same radio.C. Access Modes:Two ways of multi-access inZigBee protocol, are Beacon and Non-beacon. In non beacon enabled network, every node in the network can send the data when the channel is free. In beacon enabled network, nodes can only transmit in predetermined time slots. Here PAN coordinator allocates guaranteed time slots (GTS) for each device; therefore devices will transmit their data during their own slot. All devices should be synchronized for this process. This will be achieved by sending beacon signal. The coordinator is responsible to transmit beacon signals to synchronize the devices attached to it [4]. Network in which the coordinator does not transmit beacon signal is known as non-beacon network. It cannot have GTS and contention free periods, because the devices are not synchronized. Battery life is better than beacon enabled network, because the devices are wake up less often.ZIGBEE Protocols StackProtocol architecture is based on Open systeminterconnection (OSI). ZigBee builds on IEEE standard 802.15.4 which defines the physical and media access control (MAC) layers.ZigBee alliance defines the network layer andapplication layer. Fig.2 shows protocol stack of ZigBeesystem.Figure 3. ZigBee Protocol StackA. Physical Layer: The physical layer of the IEEE802.15.4 standard is the closest layer to the hardware, which control and communicate with the radio transceiver directly. It handles all tasks involving the access to the ZigBee hardware ,including initialization of the hardware, channel selection ,link quality estimation, energy detection measurement and clear channel assessment to assist the channel selection. Supports three frequency bands, 2.45GHz band which using 16 channels, 915MHz band which using 10 channels and 868MHz band using 1 channel. All three using Direct Spread Spectrum Sequencing (DSSS) access mode.Parameters/frequency 868Mhz 915Mhz 2450Mhz Channels 1 10 16 Data rate 20Kbps 40Kbps 250Kbps Applicability Europe USA WorldB. MAC Layer: This layer provides interface between physical layer and network layer. This provides two services; MAC data services and MAC management service interfacing to the MAC sub Layer Management Entity (MLME) Service Access Point called (MLME-SAP). The MAC data service enables the transmission and reception of MAC Protocol Data Units (MPDUs) across the PHY data service. MAC layer is responsible for generating beacons and synchronizing devices to the beacon signal in a beacon enabled services. It is also performing association and dissociation function. It defines four frame structures, are Beacon frame, Data frame, Acknowledge frame, MAC command frame. Basically there are two types of topology; star and peer to peer. Peer to peer topology can take different shapes depends on its restrictions. Peer to peer is known as mesh, if there is no restriction. Another form is tree topology. Interoperability is one of the advantages of ZigBee protocol stack. ZigBee has wide range of applications, so different manufacturer provides ZigBee devices. Z igBee devices can interact witheach other regardless of manufacturer (even if the message is encrypted).C. Network Layer: Network layer interfaces between application layer and MAC Layer. This Layer is responsible for network formation and routing. Routing is the process of selection of path to relay the messages to the destination node. This forms the network involving joining and leaving of nodes, maintaining routing tables (coordinator/router), actual routing and address allocation. ZigBee coordinator or router will perform the route discovery. This layer Provides network wide security and allows low power devices to maximize their battery life. From the basic topologies, there are threehttp: // (C) International Journal of Engineering Sciences & Research Technology[2733-2738]network topologies are considered in IEEE802.15.4 arestar, cluster tree and mesh.D. Application Layer: The application Layer is thehighest protocol layer and it hosts the application objects.ZigBee specification separates the APL layer into threedifferent sub-layers: the Application Support Sub layer, the ZigBee Device Objects, and Application Frameworkhaving manufacturer defined Application Objects.1) The application objects (APO) : Control and managesthe protocol layers in ZigBee device. It is a piece ofsoftware which controls the hardware. Each applicationobjects assigned unique end point number that otherAPO’s can use an extension to the network deviceaddress to interact with it [6]. There can be up to 240application objects in a single ZigBee device. A ZigBeeapplication must conform to an existing applicationprofile which is accepted ZigBee Alliance. Anapplication profile defines message formats andprotocols for interactions between application objects.The application profile framework allows differentvendors to independently build and sell ZigBee devicesthat can interoperate with each other in a, givenapplication profile.2) ZigBee Device Object: The key definition of ZigBeeis the ZigBee device object, which addresses three mainoperations; service discovery, security and binding. Therole of discovery is to find nodes and ask about MACaddress of coordinator/router by using uncast messages.The discovery is also facilitating the procedure forlocating some services through their profile identifiers. So profile plays an important role. The security services in this ZigBee device object have the role to authenticate and derive the necessary keys for data encryption. The network manager is implemented in the coordinator and its role is to select an existing PAN to interconnect. It also supports the creation of new PANs. The role of binding manager is to binding nodes to recourses and applications also binding devices to channels [5].3) Application support sub layer: The ApplicationSupport (APS) sub layer provides an interface betweenthe NWK and the APL layers through a general set ofservices provided by APS data and management entities. The APS sub layer processes outgoing /incoming framesin order to securely transmit/receive the frames andestablish/manage the cryptographic keys. The upperlayers issue primitives to APS sub layer to use itsservices. APS Layer Security includes the followingservices: Establish Key, Transport Key, Update Device,Remove Device, Request Key, Switch Key, EntityAuthentication, and Permissions Con guration Table.4) Security service provider: ZigBee provides security mechanism for network layer and application support layers, each of which is responsible for securing their frames. Security services include methods for key establishment, key transport, frame protection and device management.E. Topologies: There are following topologies (1)Star Topology: Star topology consists of one coordinator and any number of end devices. In star topology a master slave network model is adopted where master is the ZigBee coordinator which is FFD and slave will be either FFD or RFD. ZigBee end devices are physically and electrically separated from each other end devices and pass information through coordinator. Devices can only communicate with the coordinator. This is does not provide multi-hop networking and mesh networking. (2)Cluster Tree Topology: The cluster tree topology is similar to the star topology. The difference is that other nodes can communicate with each other so that more RFD/FFDs can be connected to non-coordinator FFDs. The advantage of this topology is the possible geographical expansion of network. (3)Mesh Topology: In mesh topology, each node can communicate any other node within its range. Mesh topology is complex to maintain and beaconing is not allowed here. But it is more robust and tolerance to fault.Figure 4: topologiesZIGBEE ApplicationZigbee Alliance targets applications “acrossconsumer, commercial, industrial and governmentmarkets worldwide”. Unwired applications are highlysought after in many networks that are characterized bynumerous nodes consuming minimum power andenjoying long battery lives.http: // (C) International Journal of Engineering Sciences & Research Technology[2733-2738]Zigbee technology is designed to best suit these applications ,for the reason that it enables reduced costs of development, very fast market adoption, and rapid ROI.Airbee Wireless Inc has tied up with Radio craft AS to deliver “Zigbee-ready solutions; the former supplying the software and the latter making the module platforms .With even light controls and thermostat producers and includes big OEM names like HP ,Philips,Motorola and Intel.With Zigbee designed to enable two-way communication , not only will the consumer be able to monitor and keep track of domestic utilities usage, but also feed it to a computer system for data analysis.Futurists are sure to hold Zigbee up and says,” See I told you so”. The Zigbee Alliance is nearly 200 strong and growing, with more OEM’s signing up. This means that more and more products and even later, all devices and their controls will be based on this standard. Since Wireless personal Area Networking applies not only to household devices, but also to individualized office automation applications, Zigbee is here to stay .It is more than likely the basis of future home-networking solutions.Table 1 Application of ZigbeeComparison to Blue ToothZigbee was developed t serve very differentapplications than Bluetooth and leads to tremendous optimizations in power consumption. Some of the key differentiators are :(a) Zigbee: It has very low duty cycle, very long primary battery life ,Static and Dynamic star and mesh networks,>65,000 nodes, with low latency available, Ability to remain quiescent for long periods without communications, Direct Sequence Spread spectrum allows devices to sleep without the requirement for close synchronization.(b) Bluetooth: It has Moderate duty cycle ,secondary battery lasts same as master, very high QoS and very low, guaranteed latency, Quasi –static star networks up to seven clients with ability to participate in more than one network, Frequency Hopping Spread Spectrum is extremely difficult to create extended networks without large synchronization cost.Advantages of ZIGBEEThe main advantages include productinteroperability, vendor independence, and accessibility to broader markets. Customers can expect increased product innovation as a result of the industry standardization of the physical radio and logical networking layers. Instead of having to invest resources to create a new proprietary solution from scratch every time, companies will now be able to leverage these industry standards to instead focus their energies on finding and serving customers. the United States. This specification maintains the same usage and architecture as wired USB devices with a high-speed host-to-device connection and connects to a maximum of 127 devices. WUSB is based on a hub and spoke topology.ConclusionThe main conclusion of this Master’s thesis project is that, yes, ZigBee is a suitable base for embedded wireless development. The main reason is that development is easy and fast. ZigBee also meets the promised technical requirements. The areas that ZigBee is likely to be used in is building automation and industrial networks. The chances seem highest in the industry since ZigBee is currently the only option for such standardized wireless networks. Even though there are some competition, due to better performance, price and compliance, ZigBee is likely to dominate the home automation market as well. PC peripherals and consumers electronics are two areas that ZigBee is very unlikely to be used in, because it offers very little over the competition.“Just as the personal computer was a symbol of the '80s, and the symbol of the '90s is the World Wide Web, the next nonlinear shift, is going to be the advent of cheap sensors.”References[1].[2]"Hands-on ZigBee: implementing 802.15.4 withmicrocontrollers" Fredeady[3]ZigBee-2007 security essentials ender y¨ ukselhanne riis nielson flemming nielson informaticsand mathematical modelling, technicaluniversity of denmark richard petersens pladsbldg 321, dk-2800 kongens lyngby, Denmark[4]Shahin farahani, "ZigBee wireless networks andtransceivers"[5]H. labiod,h. afifi,c. de santis "wi-fitm,bluetooth,zig bee and wimax".[6]Wireless sensor networks: a survey on the stateof the art and the 802.15.4 and ZigBee standardspaolo baronti, prashant pillai, vince chook ,stefano chessa , alberto gotta, y. fun hu.[7]Khanh tuan le. designing a ZigBee-ready ieee802.15.4-compliant radio transceiver. chipcon,11/2004.[8]Protocols and architectures for wireless sensornetworks holger karl university of paderborn,germany andreas willig hasso-plattnerinstitute atthe university of potsdam, germany[9]Segolene arrigault, vaia zacharaki. ” Design of aZigBee magnetic sensor node” Master ofScience thesis.[10] “Part 15.4: Wireless Medium Access Control(MAC) and Physical Layer (PHY)Specifications for Low-Rate Wireless PersonalArea Networks (LR-WPANs) “SponsorLAN/MAN Standards Committee of the IEEEComputer Society.http: // (C)International Journal of Engineering Sciences & Research Technology[2733-2738]。
MEGA 4.0
MEGA4:Molecular Evolutionary Genetics Analysis(MEGA)Software Version4.0Koichiro Tamura,* Joel Dudley,*Masatoshi Nei,àand Sudhir Kumar§**Center for Evolutionary Functional Genomics,The Biodesign Institute,Arizona State University; Department of Biological Sciences,Tokyo Metropolitan University,Tokyo,Japan;àDepartment of Biology and the Institute of Molecular Evolutionary Genetics,The Pennsylvania State University;and§School of Life Sciences,Arizona State University We announce the release of the fourth version of MEGA software,which expands on the existing facilities for editing DNA sequence data from autosequencers,mining Web-databases,performing automatic and manual sequence alignment,analyzing sequence alignments to estimate evolutionary distances,inferring phylogenetic trees,and testing evolutionary hypotheses.Version4includes a unique facility to generate captions,written infigure legend format,in order to provide natural language descriptions of the models and methods used in the analyses.This facility aims to promote a better understanding of the underlying assumptions used in analyses,and of the results generated.Another new feature is the Maximum Composite Likelihood(MCL)method for estimating evolutionary distances between all pairs of sequences simultaneously,with and without incorporating rate variation among sites and substitution pattern heterogeneities among lineages.This MCL method also can be used to estimate transition/transversion bias and nucleotide substitution pattern without knowledge of the phylogenetic tree.This new version is a native32-bit Windows application with multi-threading and multi-user supports,and it is also available to run in a Linux desktop environment (via the Wine compatibility layer)and on Intel-based Macintosh computers under the Parallels program.The current version of MEGA is available free of charge at .Since the early1990s,MEGA software functionality has evolved to include the creation and exploration of sequence alignments,the estimation of sequence diver-gence,the reconstruction and visualization of phylogenetic trees,and the testing of molecular evolutionary hypotheses. The three versions of MEGA have been released,and they integrate Web-based sequence data acquisition and align-ment capabilities(fig.1)with the evolutionary analyses (fig.2),making it much easier to conduct comparative anal-yses in a single computing environment(Kumar,Tamura, and Nei2004).Over time,MEGA has come to enhance the classroom learning experience as its use by researchers, educators,and students in diverse disciplines has expanded (Kumar and Dudley2007).The fourth version(MEGA4) contains three distinct newly developed functionalities, which are outlined below.First,we have developed a Caption Expert software module that generates descriptions for every result obtained by MEGA4.This description informs the user of all of the options used in the analysis,including the data subset used(e.g.,codon positions included),the chosen option for the handling of sites with gaps or missing data,the evolu-tionary model of substitution(e.g.,DNA substitution pat-tern,uniformity of evolutionary rates among sites,and homogeneity assumption among lineages),and the methods applied for estimating pairwise distances and for inferring and testing phylogeny.The caption also includes specific citations for any method,algorithm,and software used in the given analysis.Two examples of descriptions generated by the Caption Expert are shown infigure3.The availability of these descriptions is intended to promote a better understanding of the underlying assump-tions used in analyses,and of the results produced.This is needed because MEGA’s intuitive graphical interface makes it easy for both new and expert users to conduct a variety of computational and statistical analyses. However,some users may not immediately realize the underlying assumptions and data-handling options in-volved in each analysis.Even expert molecular and popu-lation geneticists may not be able to discern all of the assumptions implied.In general,we expect a written de-scription of methods and results to be useful for students and researchers when preparing tables andfigures for pre-sentation and publication.Second,we have now added a Maximum Composite Likelihood(MCL)method for estimating evolutionary distances(d ij)between DNA sequences,which MEGA users frequently employ for inferring phylogenetic trees, divergence times,and average sequence divergences between and within groups of sequences.In this approach, the Composite Log Likelihood(CL)obtained as the sum of log likelihood for all sequence pairs in an alignment is maximized byfitting the common parameters for nucle-otide substitution pattern(h)to every sequence pair(i,j): CL5Pi;jln lðh;d ijÞ(Tamura,Nei,and Kumar2004).This approach was previously referred to as the‘‘Simultaneous Estimation’’(SE)method,because all d ij’s are simul-taneously estimated(Tamura,Nei,and Kumar2004). The MCL approach differs from current approaches for evolutionary distance estimation,wherein each distance is estimated independently of others,either by analytical formulas or by likelihood methods(independent estimation [IE]approach).The MCL method has many advantages over the IE approach.To begin with,the IE method for estimating evo-lutionary distance for each pair of sequences will often cause rather large errors unless very long sequences are used.The use of the MCL method reduces these errors con-siderably,as a single set of parameters estimated from all-sequence pairs is applied to each distance estimation.When distances are estimated with lower errors,distance-based methods for inferring phylogenies are expected to be more accurate.This is indeed the case for theKey words:selection,genomics,phylogenetics,software,cross-platform.E-mail:s.kumar@Mol.Biol.Evol.24(8):1596–1599.2007doi:10.1093/molbev/msm092Advance Access publication May7,2007ÓThe Author2007.Published by Oxford University Press on behalf ofthe Society for Molecular Biology and Evolution.All rights reserved.For permissions,please e-mail:journals.permissions@Neighbor-Joining method (Saitou and Nei 1987),as the use of the MCL distances leads to a much higher accuracy (Tamura,Nei,and Kumar 2004).Even when the topologies estimated are the same,the use of the MCL distances often gives higher bootstrap values for the estimated phy-logenetic tree compared to the use of IE distances,which is evident from the example given in figure 4A (MCL:bold,IE:italics).In addition,the IE distances are not always estimable when pairwise distances are calculated between very dis-tantly related sequences,because the arguments of loga-rithms in the analytical formulas may become negative by chance.The probability of occurrence of such inappli-cable cases increases as the number of sequences in the data increases,the evolutionary distances become larger,and the substitution pattern becomes more complex (Tamura,Nei,and Kumar 2004).The use of the MCL method eliminates this problem effectively and allows for the use of sophisticated models in inferring phylogenies from an increasingly larger number of diverse sequences.MEGA4implements the MCL approach for estimat-ing distances between sequence pairs,average distances between and within groups,and average pairs overall with their variances estimated by a bootstrap approach.Our implementation of the MCL method allows for the consid-eration of substitution rate variation from site to site,using an approximation of the gamma distribution of evolutionary rates,and the incorporation of heterogeneity of base com-position in different species/sequences.The user also has the flexibility to estimate the numbers of transition and transversion type substitutions per site separately.Natu-rally,the MCL distances can be used for inferring phylog-enies by the distance-based methods,along with the bootstrap tests of phylogenies.MEGA4implements the MCL approach under the Tamura-Nei (1993)substitution model,in which the rates of two types of transitional substitutions (between purines [a 1]and between pyrimidines [a 2])and the rate of trans-versional substitutions (b )are considered separately by taking into account the unequal frequencies of four nucleo-tides (base composition bias).The MCL estimates of the transition/transversion rate ratio have been found to be close to the true values in previous simulation experiments (Tamura,Nei,and Kumar 2004).We have employed this feature to provide users with a facility to compute the rel-ative rates of substitutions between nucleotides based on the MCL estimates of a 1,a 2,b ,and on the observed frequencies of the four nucleotides under the Tamura-Nei (1993)model (fig.3C).For ease of comparison,we have expressed these substitution rates as relative frequencies ofsubstitutionsF IG .1.—Sequence alignment editor and Web-data mining features in MEGA4.In the Alignment Explorer (A ),the integrated web browser (B )permits downloading sequences from online databases directly into the current alignment,without the need for manual cutting-and-pasting and reformatting.The DNA sequences can be translated to the corresponding protein sequences by a single mouse click (D ),and the protein sequences can be aligned by ClustalW (E )(Thompson,Higgins,and Gibson 1994)and adjusted manually by eye.Returning to the nucleotide view automatically aligns the nucleotide sequences according to the protein alignments,and DNA and protein sequence alignments can be exported in a variety of formats for use with other programs.Alignment Editor also contains facilities for editing and importing of trace data files output from DNA sequencers (C ).MEGA4software 1597between nucleotides such that the sum of all frequencies is 100(see also Gojobori,Li,and Graur 1982).Third,we have now programmed MEGA4to run on some versions of Linux through the Wine software com-patibility layer ().The first advancement alleviates the problem of performance degradation (and the need to purchase Windows emulation software)when using MEGA on Linux.Wine is neither a hardware nor a software emulator,but an open source tool that allows for the native execution of Windows applications on Linux.Our tests of MEGA4running on Linux show the display,stability,and performance to be highly satisfactory and comparable to the native Windows system (fig.4B).Fur-thermore,investigators now report MEGA4runningonF IG .2.—A collection of menus that provide access to many different data analysis options in MEGA4,including exploration of input data set (A ),estimation of evolutionary distances (B ),inferring and testing phylogenetic trees (C ),tests of homogeneity of substitution patterns and its estimation (D ),tests of selection (E ),alignment of DNA and protein sequences (F ),and the dialog box that provides users with options to select model of substitution and data sub-setting options (G).F IG .3.—The Tree-Explorer displaying a Neighbor-Joining tree of mitochondrial 16S rRNA sequences (A ),and the description generated by the Caption Expert (B ).Estimates of the relative probabilities of nucleotide substitutions for 70control-region sequences of human mitochondrial DNA sequences are shown in (C ).The gamma shape parameter (a 50.35)was estimated using the Yang and Kumar (1996)method,and the rest of the analysis details are given in (B ).It is worth noting that the Tree Explorer shown in (A )includes a high-resolution tree drawing facility that includes displaying trees in a variety of formats,with options to display/hide branch lengths as well as clade confidence labels,and re-rooting and rearranging trees,among other functionalities.MEGA4can export the drawings to graphics programs,and can export trees in Newick format for use by other programs.Furthermore,MEGA can import and draw trees from Newick format files that have been estimated by other programs (see fig.2C ).1598Tamura et al.Intel-based Macintosh computers under the Parallels pro-gram as well as it does on Windows-native personal com-puters (see Hall 2007).The Parallels program is a native solution for Macintosh computers that permits them to simultaneously run Windows and Macintosh software.We have also built support for a multi-user environ-ment,which will allow each user of the same computer to keep his/her customized settings,including file locations,window sizes,choice of genetic code table,and previously used analysis options.This feature will facilitate educa-tional and laboratory usage,where a single computer is often shared by multiple users.In conclusion,MEGA4now contains a wide array of functionalities for the molecular evolutionary analysis of data (/features.html).It is useful to note that while we are continuously adding new methods and functions to MEGA,we do not intend to make it a catalog of all evolutionary analysis methods available.Rather,it is anticipated to become a workbench for the exploration of sequence data from evolutionary perspectives.AcknowledgmentsWe thank the colleagues,students,and volunteers who spent countless hours testing the early release versions of MEGA;almost all facets of MEGA’s design and imple-mentation benefited from their comments.We thank Ms.Linwei Wu for assistance with MEGA Web site and for handling bugs,and Ms.Kristi Garboushian for edito-rial support.We thank the two reviewers for suggesting many useful text additions,which have been included in the figure 1legend and in the text.We also thank Drs.Masafumi Nozawa and Barry Hall for comments on an earlier version of this manuscript.The MEGA software project is supported by research grants from NationalInstitutes of Health (S.K.and M.N.)and from Japan Society for Promotion of Sciences (K.T.).Literature CitedGojobori T,Li WH,Graur D.1982.Patterns of nucleotide substitution in pseudogenes and functional genes.J Mol Evol.18:360–369.Hall BG.Phylogenetic trees made easy:A how-to manual.Sunderland (MA):Sinauer Associates.Kumar S,Dudley J.2007.Bioinformatics for biologists in the genomics era.Bioinformatics.10.1093/bioinformatics/btm239.Kumar S,Tamura K,Nei M.2004.MEGA3:an integrated software for Molecular Evolutionary Genetics Analysis and sequence alignment.Brief Bioinform.5:150–163.Saitou N,Nei M.1987.The Neighbor-Joining method—a new method for reconstructing phylogenetic trees.Mol Biol Evol.4:406–425.Tamura K,Nei M.1993.Estimation of the number of nucleo-tide substitutions in the control region of mitochondrial-DNA in humans and chimpanzees.Mol Biol Evol.10:512–526.Tamura K,Nei M,Kumar S.2004.Prospects for inferring very large phylogenies by using the Neighbor-Joining method.Proc Natl Acad Sci USA.101:11030–11035.Thompson JD,Higgins DG,Gibson TJ.1994.ClustalW—improving the sensitivity of progressive multiple sequence alignment through sequence weighting,position-specific gap penalties and weight matrix choice.Nucleic Acids Res.22:4673–4680.Yang Z,Kumar S.1996.Approximate methods for estimating the pattern of nucleotide substitution and the variation of substitution rates among sites.Mol Biol Evol.13:650–659.William Martin,Associate Editor Accepted May 2,2007F IG .4.—(A )Bootstrap support for the branching order of 16Laurasiatheria species reconstructed with MCL approach (bold)and without MCL approach (italics)under the Tamura-Nei (1993)model (see figure 3B for rest of the analysis details).The 16S rRNA sequences used were downloaded from GenBank and were aligned in MEGA4using CLUSTALW (accession numbers:AJ428578,NC004029,X72004,AF303109,NC008093,DQ480502,X97336,X79547,DQ534707,AJ554051,AJ554061,NC000889,NC007704,AB074968,NC005044,and NC001941).(B )Comparison of MEGA4performance benchmarks on Windows and Linux (with Wine application compatibility layer).Identical hardware configuration was used,and example data sets included in the MEGA4installation were employed.The results show that computations executed under Wine are penalized by about 2s,which is attributable to the need for Wine’s initialization.MEGA4software 1599。
ELASTIC INTERFACES
Elastic interfaces. Maritime instrumentation as an exampleProceedings of the CSAPC'99, Université de Valenciennes, Valenciennes France. 1999, 35-41.Peter Bøgh AndersenCenter for Human Computer Interaction/ University of Aarhus. E-mail: pba@imv.aau.dkABSTRACTThis paper introduces the general notion of elastic interfaces and the derived concepts of transparency and tai-lorability, and exemplifies the notions in the domain of maritime instrumentation. It concludes that maritime automation to some degree already follows principles of transparency and possibly might benefit from adopting a disciplined version of tailorability.Keywords: maritime instruments, tailorability, transparency, HCI.ELASTIC INTERFACESThis paper reports on some ongoing work in the project Elastic Systems under the Danish Center for Human Computer Interaction1.First a definition: an elastic interface is one that lets the user move continuously in two dimensions:1.The dimension of access. The dimension measures the degree to which the user has access to changingthe underlying features of the system that determine its possibilities of use. At one extreme, the user is confined to only using the system for the purpose it was designed for. At the other extreme, the user has the power of a normal programming environment. The interesting part is the middle area of this dimen-sion.2.The dimension of agency. The dimension measures the degree to which the user or the system deter-mines the course of events. At one end, the user initiates every action and the system merely responds. At the other extreme, the system handles the whole process and the user can only start and stop it. Again it is the middle part that is interesting.The term elastic can be seen as a name for an already existing tendency, noted in Gentner & Nielsen 1996, to-wards designing systems that are mutable (buzz word: tailorability) and which may exhibit variable degrees of autonomy (buzz-word: autonomous agents).The term elastic stresses that computer systems are dynamic systems that involve two main kinds of dynamics: a synchronic dynamics between user and system and a diachronic dynamics of adaptation and development. It in-vites us to view a computer system more as an evolving process than as a stable object. This paper applies the concept to the maritime domain that is characterized by two interesting features:•The domain is a safety-critical domain which creates special conditions for tailoring.•Maritime instruments by nature have an autonomous environment. The navigator cannot escape sharing con-trol over instruments with wind, waves, and currents.A simple maritime example of (1) is that some maritime systems can be toggled between sea and harbor condi-tions, displaying different types of information and enabling different actions in the two conditions. Alarms are a simple example of (2). It is a problem that maritime systems give too many and too undifferentiated alarms which the navigator must acknowledge and thereby possibly may get diverted from the real task at hand. Alarms are sounded on the initiative of the machine to call the attention of the operator and thus regulate the balance of power between the two.In the domain of process control, the access to changing the system should be strongly constrained by safety considerations. Movement on the agency scale is already carefully regulated in the maritime domain.THE CONCEPT OF TAILORINGBasically, tailoring is motivated by the inability of the designer to accurately predict the needs of the use situation, and most modern system therefore offer some kind of tailoring: setting preferences and redesigning the menu structure. End-user programming in a scripting language is not uncommon. However, there is often too large a gap between the understanding developed during usage and the concepts needed for tailoring purposes, and users are1 The center is funded by the Danish National Research Foundation. The empirical data are from three ferries, a freighter, a2 weeks cruise with a large container ship, and from two transcribed simulated voyages at the Danish Maritime Institute. It must be borne in mind that the ship types are not yet representative.often reluctant to accept this additional burden (Wasserschaff & Bentley 1997: 307). In Andersen 1997[1990]: 180 I suggested the following principle for tailorable systems:1.Experiences gained from using the system should be applicable for modifying and changing it.and later research in tailorability has emphasized the same point (Malone, T. W. K-Y. Lai & C. Fry. 1995: 178). In this paper I shall explore the following very simple and general principle:2.The geometry of tailorable artifacts should be self-similar.By self-similarity I mean the mathematical property that parts of the object are a scaled down version of the whole object (see Peitgen, Jürgens & Saupe 1992 for a good treatment of the concept and many examples). Self-similar objects have the property that if a part is enlarged it displays the same morphology as the whole object. Applied to computer systems it means that the internal structure of a module of the system is similar to the structure of the whole system. This enables experiences suitable for handling the whole system to be applicable for understanding the details of the parts too. Two specific criteria can be deduced from this. One pertains to understanding and one to acting:3.The Principle of Transparency: What goes on between two internal objects of the system is analogous towhat goes on between the user and the interface-objects.4.The Principle of Tailorability: Changing internal objects of the system is like changing the interface-objects of the system.WHY TRANSPARENCY AND TAILORABILITY?In this section I shall present two arguments for transparency and tailorability in systems for maritime navigation. The transparency argument concerns situations of malfunction of machinery, whereas the tailorability argument takes its point of departure in the fact that navigation is divided into phases and therefore causes the information needs to change.Transparency: Handling and Preventing MalfunctionsWhen I visited a modern ferry with integrated bridge and many useful systems for aiding the captain, the two navigators pulled me aside when I was leaving: they were quite satisfied with their equipment but they wanted to voice one worry: how do we know what is left when a component falls out?With the old equipment it was easy to figure out what was left if, for example, the GPS (Global Positioning System) stopped functioning. The GPS was an isolated instrument that showed the ship’s position, but had no interconnections with the other instruments. So the captain knew “what was left” when the GPS malfunctioned. This is much more difficult to assess in integrated bridges, where components deliver data to many other com-ponents. Thus, it is important that the user has a chance to understand the nature of malfunctions and figure out how the modules interact with each other.Even more important is the ability of acting in the case of malfunction. If the automatic steering system breaks down, the navigator should be able to easily revert to manual steering, because the crew cannot stop work because of an instrument failure, as the user of a safe PC can do.Prevention of malfunctions is another good reason for transparency. For example, in Maersk 1999 it is required that instruments and steering gear are regularly checked for correctness and functionality. If such control is to be possible at all, the workings of sensors and actuators must be understandable.But creating such understanding is no trivial problem: if the automatic systems are “strong, silent, clumsy, and difficult to direct” (Woods 1996: 6) they generate confusion and errors.Tailorability: PhasesA main argument for tailorability is the fact that navigation falls into distinct phases, each with their own infor-mation needs.There are three main phases: navigation in open waters, coastal waters and harbor2. The phases are clearly marked by different types of cooperation and information needs. For example, in open waters one officer can han-dle the navigation, whereas two to four may be needed in the two other cases.In one ship I visited there was the following sequence that reflects a decreasing curve of tension: (1) When things are difficult inside the harbor, the captain handles the steering advised by the pilot and helped by the chief officer. (2) Then the helmsman is called in and receives commands from the captain. (3) Outside the trafficked 2 In production plants, phases can be found e.g. in the start-up procedure.area, the pilot leaves, and the watch officer replaces the chief officer. (4) The helmsman leaves, and (5) in open waters (in the daytime and in clear weather) only the watch officer is on the bridge. The sequence 1⇒ 5 is used when leaving the harbor, whereas the reverse 5 ⇒ 1 is used during entry.Not only the work organization, butalso the place of work changes with thephases. During berthing, navigation canbe moved to the bridge wing so that thenavigator can see the ship’s side. Eachbridge wing must therefore contain acopy of the relevant displays and con-trols.Information needs change as well. Out-side the harbor, the radar is used to plotbearings and distances to other ships, butinside there may be too many vessels tooclose for this to be useful. In open wa-ters, the GPS information of latitude andcoastal waters. Inside a harbor, the importance of visual assessments ofdistances increases; some instruments increase their relevance as well,while others change their function. For example, drift and set become im-portant during berthing. The radar is now used for measuring the distanceto the quay, and the V(oyage) M(anagement) S(ystem) for assessing theposition of the ship in the basin. Other instruments become obsolete: forexample engine information about revolutions per minute (rpm)3, fins, andautopilot have nothing to do in the bridge wing that is specialized forberthing maneuvers. This is illustrated in Fig. 1. that displays the occur-rences of conversations dealing with distance and positions from a simu-lated voyage at the Danish Maritime Institute (see Andersen 1998b).With a metaphor, one may say that while navigating a ship is like drivinga train in open sea, it is more like riding a horse inside the harbor.These shifts of information needs are not reflected in instrumentation. Forexample, for some strange reason, many instruments give latitude and lon-gitude a prominent place in the display. In the conning display in Fig. 2, itoccupies the “best” place in the top half of the display. However, the vesselwas a ferry that always sailed in restricted waters and the numbers were never used (May 1999).In cases like this, one might suspect that some kind of tailoring of the interface would be appropriate, allowing the captain to select the information relevant for the phase he is in and give it a prominent place.This would not only benefit the person using the instruments, but also the other officers he is cooperating with since good bridge practice requires all to verbalize their actions so that everybody knows what goes on. This abil-ity is taught on courses, is encouraged in ship owner's guidelines (Maersk 1999) and can in fact be observed in good captains:(H, verbalizing)I am down on ninety nine, a hundred now, ninety nine, or ninety eight (E, volunteering information)We are to start the turn when we have the buoy to the northFor more examples, see Andersen 1998a, Section 2. This kind of discrete monitoring and helping out is not unique to maritime navigation but is probably common in all kinds of process control, cf. Heath & Luff 1992 that con-clude that in order for such mutual monitoring to occur, the size and location of the display is crucial.There is therefore reason to believe that giving the phase-relevant information a prominent position on the bridge may enhance the mutual monitoring and help.AGENCY AND ACCESS IN EXISTING TOOLSIn this section we shall see how the needs for transparency and tailorability are realized in existing instrumenta-tion.3 Although the exact rpm number is irrelevant, there is still a need for a more coarse engine feed back telling the captain whether his command is being executed: does the ship move in the right direction and do the revolutions go towards what he has ordered (e.g. full ahead).AgencyAgency is regulated in two ways, namely•Paradigmatic (either-or): Either the system or the user is doing the task, but not both at the same time. The system take over tasks which the navigator will perform under other circumstances.•Syntagmatic (both-and): Both system and user are active in performing the task, the system allowing the navigator to manipulate work objects of different granularity.Voyage Management Systems exemplify the paradigmatic type: they allow the navigator to draw a track the ship should follow, and when a plan is executing, the system replaces the navigator as the controller of the rudder. Integrated joysticks are examples of syntagmatic automation. On one ship, a joystick controlled the water-jets of the ship in unison, so that the navigator did not have to think about manipulating the individual jet, but could imagine that he manipulated the whole ship. The automation did not replace the navigator, but allowed him to work on a higher level of granularity.Because of its safety critical nature, maritime instrumentation is based on the principle that if higher level auto-mation falls out, there must always be a lower level that can be used. Steering is a good example. In one ship there were at least six following levels: VMS, Autopilot, Helm, N(on)F(ollow)U(p), Manual steering of rudder ma-chine, Physical movement of rudder by means of wires (Figs. 3-5). Similarly, instead of the integrated joystick the navigator could use individual sticks controlling the individual jets, and he could even manipulate their hydraulics directly.Fig. 3. VMS system on bridge Fig. 4. Autopilot on bridge Fig. 5. Manual steering of ruddermachine in engine room.Shifts of agency is carefully regulated in both instrumentation and operating guidelines.AccessWhereas shifts of agency play a major role in ship design, shifts of access are not so pronounced. However, elec-tronic charts can be tailored, so that only a selection of the available buoys, beacons, lights, depth contours, cau-tionary areas, and currents are displayed. The purpose is to avoid cluttering the chart. Another area is configuring instruments.USING TRANSPARENCY AND TAILORABILITY IN DESIGN OF MARITIME INSTRUMENTSIn this section I shall present some ideas for exploiting transparency and tailorability more systematically in the maritime domain.TransparencySome existing systems tend to conform partially to the principle of transparency. The higher level automation handles the lower level in the same way as the user does. For example, both VMS and user can input course or-ders to the autopilot, both autopilot and user can input rudder orders to the rudder system, and both rudder system and user can control the hydraulics of the rudder machine. Thus, the system is self-similar with respect to prag-matics, i.e. the user-system relation. I can understand the workings of the VMS in analogy to the way I myself interact with the autopilot.One could consider extending the pragmatic self-similarity to semantic self-similarity. This would mean that the internal system structure is based on the concepts the navigator uses when manipulating the whole system. The purpose is to make it easier for the user to move along the dimension of agency. Using action-based intentional concepts for process control has already been suggested in Lind 1990, 1994. In order to achieve semantic self-similarity, the action schemata must be extracted from the language pertaining to the usage of the instruments, andin the maritime domain this means maneuvers . The following schema is extracted from an analysis of typical ma-neuvers (Andersen 1998a).1.Goal: what we want to achieve, e.g. avoid collision with foreign ship.2.Means: actions that lead to the goal, e.g. using the rudder.3.Help: forces that enhance the means, e.g. exploiting the wind.4.Obstacles: forces that hinder the goal, e.g. low speed.5.Negative side-effects: negative state caused by the means, e.g. the turning moment produced during a crash stop achieved by putting the engine in reverse.6. Countermeasures: actions that aim at eliminating obstacles and negative side-effects, e.g. compensatingwith rudder in (5) or increasing speed in (4).For a system to be semantically self-similar its internal architecture must be based on concepts like these. If this is the case, the navigator can use his skills from maneuvering for understanding and repairing the system.TailorabilityWe cannot tailor collective, safety critical systems like a ship bridge in the same way as we do with our personal PC. Tailoring must be controlled, since it will not do that an officer has to look for the echo sounder in a new place each time he takes over the watch. Still, fixed and hardwired instruments do present problems for a work domain that is divided into phases. What is useful in one phase, is irrelevant for another. Furthermore, a panel that gives longitude and latitude a prominent position is well suited if the vessel is ocean-going, but becomes irrelevant if it is turned into a ferry.Here is a possible scenario: sensor signals are dissociated from individual displays, and the hardwired displays are replaced by a panel of screens that can be treated logically as one screen so that is possible to move interface objects from one screen to another. Some interface objects, such as the VMS and the radar, are produced by manufacturers, whereas others objects can be designed by ship-owner, captain, or officers according to a well-defined hierarchy of constraints. For example, maritime law will rule that some instruments are obligatory and must have a certain size.Judging from observations, the envisioned tool should offer facilities such as: moving and scaling objects, com-bining objects; changing mode of display. I conclude by two authentic ex-amples that illustrate this.1. An officer of the watch had forgotten that the rudder limit and R(ate)O(f) T(urn) limit were set and could not figure out why the wheel did notproduce the desired turning. The reason was that the ROT display and thelimits were represented in different places. In some autopilots limits areknobs whose position is difficult to see in the dark (see Fig. 4).After having heard the story, another officer produced on the spot Fig. 7 asa solution to the problem. In his redesign he combined four old displays into one: the rudder angle and ROT indicators, and the two knobs from theautopilot. The curved bars indicate the angles of the rudder and rate of turn,and the bullets the limits of both. By looking at the new display the officer ofthe watch can immediately see the correct cause of the failure.To realize this design, the officer needs facilities for changing the repre-sentation from the original digits to bars, and for combining four stand-aloneinstruments into one.2. A captain wants to use fuel most economically, i.e. to achieve most speed for the lowest fuel consumption. On the Maihak display in front of him (Fig. 8) he sees 3 numbers on a display:revolutions per minute (rpm), torque, and horse power.These numbers express the power produced by the engine. The captainwants to hit the most economical way of producing speed, so he wants tocompare the three numbers with the speed indicator placed some distanceaway.This is difficult two reasons: he must remember combinations of the fourfigures to hit upon the most economical one and his eye must move fromone place in the panel to another one. Therefore, instead of looking at thenumbers in front of him, he turns and follows the development of the rpm and speed on trend curves available on a screen behind him.The curves are better in two respects: the rpm and speed curves are drawnin the same diagram and therefore it is easy to compare them and see whenFig. 8. The Maihak display with rpm, torque and horse power.they increase in parallel and when speed starts to increase less than rpm. In addition, he does not need to remem-ber past values, because this is what trend curves do.To do this in reality the captain needs facilities for changing numbers into trend curves and combining two dis-plays (rpm and speed) into one.The examples are simple in the sense that they can be realized by changing representations and combining dis-plays. However, a tailorable bridge could also stimulate imagination of the officers and give rise to new displays that the landlubbers of the instrument manufacturers would never think of.REFERENCESANDERSEN, P. BØGH (1997 [1990]). A theory of computer semiotics. Semiotic approaches to construction and assessment of computer systems. Doctoral dissertation. Cambridge University Press.ANDERSEN, P. BØGH (1998a). Analysis of Maritime Operations II. Center for Human-Machine Interaction. Report CHMI-2-1998. http://www.chmi.dk/ANDERSEN, P. BØGH (1998b). Analysis of Maritime Operations III. Center for Human-Machine Interaction. Report CHMI-3-1998. http://www.chmi.dk/CARROLL, J. & M. ROSSON (1987). Paradox of the Active User. In J. Carroll (ed.): Interfacing Thought. Cambridge, MA: MIT Press, 80-111.DMI (1997). Manoeuver Theory.GENTNER, D. & J. NIELSEN (1996). The Anti-Mac Interface. CACM 8: 70 - 82.HEATH, C. & P. LUFF (1992). Collaboration and control. Crisis management and multimedia technology in London under-ground line control rooms. Computer Supported Work 1: 69-94.HUTCHINS, E. (1996). Cognition in the wild. Cambridge, Mass.: The MIT Press)LIND, M.(1990). Representing goals and functions of complex systems. Institute of Automatic Control Systems. DTU. 90-D-381.LIND, M. (1994). Modeling goals and functions of complex industrial plants. Applied Artificial Intelligence 8: 259-283. MACKAY, W. (1991). Triggers and Barriers to Customizing Software. In Proc. of CHI’91, New Orleans, Louisiana, 1991.New Work: The ACM Press, 153-160.MALONE, T. W. K-Y. LAI & C. FRY. (1995) Experiments with Oval: A Radically Tailorable Tool for Cooperative Work.ACM Transactions on Information Systems, 13/2: 177-205.MAY, M. (1999). Cognitive aspects of interface design and human-centered automation on the ship bridge: the example of ARPA/ECDIS integration. Presentation for PIC’99, Bath, June 21.23.MAERSK (1999). Guidelines for navigators. Maersk Vessels. Technical organisation, level 2 document. NAU610. Version April, 1999PEITGEN, H.O, H. JÜRGENS & D. SAUPE (1992). Chaos and fractals. Berlin: Springer-Verlag.WASSERSCHAFF, M. & R. BENTLEY (1997). Supporting Cooperation through Customisation: The Tviews Approach.Computer Supported Cooperative Work 6: 305-325.WOODS, D. D. (1996). Decomposing automation: apparent simplicity, real complexity. In: R. Paraduramen & M. Mouloua (eds.): Automation and human performance: theory and applications., 1-17. Mahwah, New Jersey: Erlbaum.。
动态信任关系建模和管理技术研究
49.Gerck E Toward Real-World Models of Trust 1998
20.Chen R.W.Yeager Poblano:A Distributed Trust Model for Peer-to-Peer Networks.Sun Microsystems 2000
21.Winsborough W H.N Li Towards Practical Automated Trust Negotiation 2002
11.A.Abdul-Rahman.S.Halles A distributed trust model 1997
12.A.Abdul-Rahman.S.Hailes Using recommendations for managing trust in distributed systems 1997
34.F.Cornelli.E.Damiani.S.C.Vimercati.S.Paraboschi,P.Samarati A reputation-based approach for choosing reliable resources in peer-to-peer networks 2002
●由于PeerTrust和SecDWTrust都采用了社群信任查询的方法,因此在可信节点占大多数的情况下可以避免虚假的推荐,在~定程度上减轻了联合作弊的危害。而DwTrust 仪依赖直接经验进行判断,在缺乏赢接交互历史的情况下,无法识别联合欺骗行为。
Journal of materials science letter例稿
LETTERActivation enthalpies for oxygen ion motion in cubic yttria-stabilized zirconiaR.J.Darby ÆR.V.KumarReceived:16June 2008/Accepted:2September 2008/Published online:19September 2008ÓSpringer Science+Business Media,LLC 2008IntroductionThe high oxygen ion conductivity of the yttria-stabilized zirconia (YSZ)system is well known.The conductivity at 1000°C increases with YO 1.5addition to around 15–18mol.%YO 1.5,when the dopant content is just sufficient to fully stabilize the cubic fluorite phase.Further yttria additions result in a decrease in the conductivity [1,2].There are several contributory factors suggested to explain the conductivity decrease:formation of defect clusters reducing oxygen vacancy mobility [3–5],hindrance of oxygen ion motion due to the increasing presence of ‘large’yttrium ions [6–8],and the increasing presence of a grain boundary phase [9].Support for the influence of defect clusters is the observation that the activation enthalpy for oxygen ion conduction decreases at high temperatures (decreasing by *0.2eV above *650°C for 15mol.%YO 1.5)[4,10,11].This has been correlated with the breaking-up of short-range order [11].Although there is a wide compositional range of cubic stability,only a few studies have experimentally investi-gated yttria contents above 33mol.%YO 1.5[12–15].In these studies the activation enthalpy was observed to increase with yttria content until *33mol.%YO 1.5where the activation enthalpy ‘saturated’to a value of *1.3eV.The authors are unaware of any comment made so far on this saturation effect and it is thus the focus of this letter,allowing an insight into the different contributions to the conductivity decline.Powders containing between 21and 53mol.%YO 1.5were prepared via roll-milling 8mol.%Y 2O 3-stabilized zirconia powder (TZ-8Y,Tosoh Corporation)with addi-tional yttria powder (99.99%purity,Sigma–Aldrich)in ethanol for 24h.These powders were subsequently cal-cined at 900°C for 4h,pressed into cuboids or pellets and then sintered at 1500°C for 48h in air.Platinum elec-trodes (Platinum Ink 6926,Metalor)were applied to the faces of the samples for the conductivity measurements.All samples were confirmed to possess the cubic fluorite structure via X-ray diffraction and Raman spectroscopy and had relative densities of over 95%.Both ac impedance and 4-point dc measurements were made using a Solartron 1260frequency response analyser coupled with a Solartron 1287potentiostat.When possible,the conductivity of each of the samples was measured by the 4-point dc method between *480and 1000°C (Fig.1).The variation of conductivity with composition agrees with previous work,decreasing with an increasing yttria content.The variation of conductivity,r ,with temperature,T ,is given in an Arrhenius form by:ln r T ¼D H Ãk B Tþln A ;ð1Þwhere D H *is the activation enthalpy,k B the Boltzmann constant and A is known as the pre-exponential ing the gradients in Fig.1along with (1)it is possible to determine the activation enthalpy for each sample.As previously stated,it is known that the Arrhenius plots display a temperature-dependent gradient for low yttria contents.Temperature-dependent gradients have previ-ously been observed up to *38mol.%YO 1.5[16].All the samples measured here display such a temperature-depen-dent gradient.This is illustrated by the highest yttriaR.J.Darby (&)ÁR.V.KumarDepartment of Materials Science and Metallurgy,University of Cambridge,Cambridge,UK e-mail:rjd44@J Mater Sci (2008)43:6567–6570DOI 10.1007/s10853-008-2983-5content sample tested,53mol.%YO 1.5(Fig.2).Although two linear fits have been used to analyse the data,it must be emphasized that this does not necessarily indicate that there are two distinct regions of behaviour with an intermediate transition temperature.Such analysis,using two linear fits,has been common in the past [4,10,11],but recent work has clearly illustrated the activation enthalpy continuously changing with temperature [16].Activation enthalpies for all samples were calculated in the lowest temperature range where data were obtained for all samples,between 610and 675°C,as well as the highest temperature region measured,between 890and 1000°C (Fig.3).The difference between the low temperature andhigh temperature activation enthalpies is remarkably con-stant,varying between 0.12and 0.18eV.As the activation enthalpy decrease has been correlated with the removal of short-range order in the low yttria content samples,it can be suggested that these higher yttria content samples are undergoing a similar dissociation of defect clusters [11].This results in the similar reduction in activation enthalpy.The activation enthalpy for oxygen ion motion is com-posed of two contributions;a migrational enthalpy,D H m ,and an association enthalpy,D H a [17]:D H üD H m þD H a :ð2ÞThe migrational enthalpy represents the energy barrier between adjacent oxygen ion sites while the association enthalpy accounts for such interactions that introduce short-range ordering (such as between the relatively negatively charged dopant cation and positively charged oxygen vacancy as well as the lattice relaxation accommodating the differently sized dopant).These interactions must be overcome to mobilize oxygen ions and vacancies for conduction.Variation in either the migrational enthalpy,association enthalpy or both may account for the compositional variation in the activation enthalpy observed here.Lee and Navrotsky [18,19]have recently reported enthalpy of formation data for the YSZ system that will aid the interpretation of the relative contributions of the dif-ferent enthalpies in (2).From these data it is possible to calculate the enthalpy of mixing,D H mix ,which is the deviation from a linear variation between the end-members (Fig.4).The end-members are the enthalpy for the tran-sition between the stable low-temperature monoclinicphase of zirconia and cubic fluorite zirconia,D H ZrO 2m !c ;and the transition of the stable low-temperature C-type cubicyttria to cubic fluorite yttria,D H YO 1:5C !c :Thenegativedeviation from the linear line suggests favourable interac-tions.Such interactions are responsible for the association enthalpy term in(2).In order to understand how the variation in the enthalpy of mixing is affecting the interactions experienced by each oxygen vacancy,the data are normalized relative to the number of oxygen vacancies present(Fig.5).This shows that additional oxygen vacancies experience a similar amount of interactions below*30mol.%YO1.5,repre-sented by a relatively constant enthalpy of mixing per oxygen vacancy with increasing yttria content.Thus it can be inferred that the association enthalpy term in(2)is constant below*30mol.%YO1.5.The activation enthalpy variation in this composition range must be entirely due to an increasing migrational enthalpy with yttria content.An increase in migrational enthalpy would be expected with increasing yttria content as it results in an increasing presence of larger yttrium ions,sterically hindering the motion of oxygen ions(the zirconium and yttrium ions have ionic radii of0.84and1.019A˚,respectively[20]). Thus the activation enthalpy is initially increasing with yttria content due to the increased likelihood of a mobile oxygen ion having to pass a large yttrium ion.This observation of a‘blocking effect’is in agreement with computer simulations[6–8].In particular it corroborates the conclusions of a recent analytical model of YSZ con-ductivity by Martin[21].Martin found that the variation in conductivity with dopant content was best modelled by a constant favourable enthalpy of association coupled with an increasing amount of blocking caused by the increasing proportion of large yttrium ions.This is exactly what has been experimentally observed here for samples with yttria contents below*30mol.%YO1.5.Beyond*30mol.%YO1.5,the enthalpy of mixing per oxygen vacancy becomes less exothermic,suggesting fewer favourable interactions(possibly due to the increasing presence of yttrium ions and oxygen vacancies beginning to interact repulsively,coupled with increasing amounts of strain).This indicates that the association enthalpy contribution in(2)is becoming less with further yttria additions.The migrational enthalpy term will keep increasing due to the increasing presence of relatively large yttrium ions.An increasing D H m and a decreasing D H a result in the almost constant activation enthalpy observed in this composition range.A further increase in the activation enthalpy is observed for the53mol.%YO1.5sample.As the association enthalpy is continuing to decrease this must be due to a greater increase in the migrational enthalpy as more than half the cation sites are now occupied by large yttrium ions.Overall the activation enthalpy variation with yttria content has been explained in terms of the migrational and association enthalpy contributions.At low yttria contents (below*30mol.%YO1.5)the activation enthalpy increases due to an increasing migrational enthalpy term, caused by the increasing presence of large yttrium ions. The association enthalpy is constant in this region as each defect cluster is relatively isolated,the additional oxygen vacancies created on yttria addition are thus in similar environments and are similarly‘associated’.At high yttria contents(beyond*30mol.%YO1.5)the defect clusters are no longer isolated and begin to interact,becoming destabilized and reducing the association enthalpy.The migrational enthalpy keeps increasing resulting in the sat-uration of the activation enthalpy observed.As the activation enthalpy is the dominant factor affecting the measured conductivity value[22],this anal-ysis suggests that the sudden conductivity decreasebeyondthe yttria content required for cubic stabilization is caused by an increasing amount of large yttrium ions in the zir-conia matrix.It is not caused by an increasing amount of defect cluster formation.For completeness the suggestion that the conductivity decreases due to the increasing pres-ence of a grain boundary phase was investigated by impedance spectroscopy.No such additional contribution to the conductivity at higher yttria contents was observed. Acknowledgements R.J.D.and R.VÁK.would like to thank Ian Farnan for helpful discussions.R.J.D.acknowledges funding from UK Engineering and Physical Sciences Research Council.References1.Badwal SPS(1992)Solid State Ionics52:23.doi:10.1016/0167-2738(92)90088-72.Hattori M,Takeda Y,Sakaki Y,Nakanishi A,Ohara S,Mukai Ket al(2004)J Power Sources126:23.doi:10.1016/j.jpowsour.2003.08.0183.Kilner JA,Steele BCH(1981)In:Sørensen OT(ed)Masstransport in anion-deficientfluorite oxides,1st edn.Academic Press,New York4.Badwal SPS(1984)J Mater Sci19:1767.doi:10.1007/BF005502465.Ioffe AI,Rutman DS,Karpachov SV(1978)Electrochim Acta23:141.doi:10.1016/0013-4686(78)80110-86.Pornprasertsuk R,Ramanarayanan P,Musgrave CB,Prinz FB(2005)J Appl Phys98:103513.doi:10.1063/1.21358897.Meyer M,Nicoloso N,Jaenisch V(1997)Phys Rev B56:5961.doi:10.1103/PhysRevB.56.59618.Shimojo F,Okazaki H(1992)J Phys Soc Jpn61:4106.doi:10.1143/JPSJ.61.41069.Yamamura H,Utsunomiya N,Mori T,Atake T(1998)Solid StateIonics107:185.doi:10.1016/S0167-2738(97)00534-110.Filal M,Petot C,Mokchah M,Chateau C,Carpentier JL(1995)Solid State Ionics80:27.doi:10.1016/0167-2738(95)00137-U 11.Gibson IR,Irvine JTS(1996)J Mater Chem6:895.doi:10.1039/jm996060089512.Strickler DW,Carlson WG(1964)J Am Ceram Soc47:122.doi:10.1111/j.1151-2916.1964.tb14368.x13.Irvine JTS,Feighery AJ,Fagg DP,Garcia-Martin S(2000)SolidState Ionics136:879.doi:10.1016/S0167-2738(00)00568-3 14.Dixon JM,LaGrange LD,Merten U,Miller CF,Porter JT II(1963)J Electrochem Soc110:276.doi:10.1149/1.2425731 15.Irvine J,Gibson I,Fagg D(1995)Ionics1:279.doi:10.1007/BF0239020816.Weller M,Herzog R,Kilo M,Borchardt G,Weber S,Scherrer S(2004)Solid State Ionics175:409.doi:10.1016/j.ssi.2003.12.044 17.Goodenough JB(2003)Annu Rev Mater Res33:91.doi:10.1146/annurev.matsci.33.022802.09165118.Lee TA,Navrotsky A,Molodetsky I(2003)J Mater Res18:908.doi:10.1557/JMR.2003.012519.Navrotsky A,Benoist L,Lefebvre H(2005)J Am Ceram Soc88:294220.Shannon RD(1976)Acta Crystallogr A32:751.doi:10.1107/S056773947600155121.Martin M(2006)J Electroceram17:765.doi:10.1007/s10832-006-6007-z22.Almond DP,West AR(1987)Solid State Ionics23:27.doi:10.1016/0167-2738(87)90078-6。
- 1、下载文档前请自行甄别文档内容的完整性，平台不提供额外的编辑、内容补充、找答案等附加服务。
- 2、"仅部分预览"的文档,不可在线预览部分如存在完整性等问题,可反馈申请退款(可完整预览的文档不适用该条件!)。
- 3、如文档侵犯您的权益，请联系客服反馈,我们会尽快为您处理(人工客服工作时间：9:00-18:30)。
Reference Notes to Slides•Many of the references cited in this itemized list,keyed to slide numbers,can be found in the list of references that follows. This longer list was prepared for another purpose.•Slide2Einstein quotes from web search.•Slide3adapted from Shakespeare,Julius Caesar.•Slide11,see reference to Krantz.•Slide14,from web search.•Slides19,20,21,see reference to Neumann.•Slide25•Slides26,27,28,see Jauch and Liboff.•Slide30,see Ambegaokar.•Slides33–36:Hoeffding’s inequality is available from many sources,including Fine[2006].•Slides38,39,40are quotations from John Venn,cited below.•Slide42is available from many sources,including Fine[1973].•Slides43,44issues addressed in Kaplan[1977].•Slides46–49based on prior research of Walley and Fine[1982] and Kumar[1982].•Slides51,52drawn from Mises[1957,1981]although the ideas are much earlier.•Slide53from Jeffrey[1992].•Slide54from Weinberg[1989].•Slides56–60,primarily from Walley and Fine[1982].•Slide61based on work of Fine,Fierens,Rego as reported in Fierens[2003,2007]and Rego[2005,2007].•Slides63–68see Sadrolhefazi and Fine[1994]for the main re-sults and citations to previous work by Grize,Kumar,and Pa-pamarcou.•Slides70–87motivated by Keynes[1921]and uses of Kolmogorov complexity to define induction as reported in Li and Vit´a nyi [1997]and based on1960s work by Ray Solomonoff,Gregory Chaitin,and Kolmogorov.This work is offered tentatively.•Slides89–95reflects long-held thoughts of the author on the limitations of subjective probability.1References[1]Ambegaokar,Vinay[1996],Reasoning about Luck:probability and its uses inphysics,Cambridge University Press,Cambridge.[2]Bernoulli,Jacob[1713],Ars Conjectandi,Pars Quarta.Reputed to have beencompleted ca.1692.Translated from the Latin by Bing Sung,Dept.of Sta-tistics,Harvard University,Feb.1966.See also Hacking(1975)for much on Bernoulli.[3]Bogdan,Radu[1982],Henry E.Kyburg,jr.&Isaac Levi,Bogdan is the editor,D.Reidel Pub.,Dordrecht–Holland,Boston,MA[4]Born,Max[1926],Zur quantenmechanik der Stoßvorg¨a nge,Zeitschrift furPhysik,XXXVII,865.[5]Carnap,Rudolf[1952],The Continuum of Inductive Methods,University ofChicago Press,Chicago.[6]Carnap,Rudolf[1962],Logical Foundations of Probability,second edition,Univ.of Chicago Press,Chicago.[7]Cozman,Fabio,Lonnie Chrisman[1997],Learning Convex Sets of Probabilityfrom Data,CMU RI TR9725,Carnegie Mellon Univ.,Pittsburgh,PA. [8]Fierens,Pablo I.[2003],Towards a Chaotic Probability Model for FrequentistProbability,Ph.D.dissertation,Cornell University,Ithaca,NY.[9]Fierens,Pablo I.[2007],An extension of chaotic probability models to real-valued variables,Gert de Cooman,Jirina Vejnarova,Marco Zaffalon,eds., 5th Inter.Symp.on Imprecise Probability:Theories and Applications,Prague, 193–202.[10]Fine,Terrence L.[1973],Theories of Probability:An examination of founda-tions,Academic Press,NY.Copyright has since reverted to the author. [11]Fine,Terrence L.[1975],An argument for comparative probability,in RobertButts and Jaakko Hintikka Basic Problems in Methodology and Linguistics, Part Three,D.Riedel,Dordrecht.Proceedings of the Fifth International Con-gress of Logic,Methodology,and Philosophy of Science.[12]Fine,Terrence L.[1983],Foundations of probability,in editors-in-chief SamuelKotz,Norman L.Johnson;associate editor,Campbell B.Read,Encyclopedia of Statistical Sciences,3,John Wiley&Sons,NY.(volumes produced in the years1982–1989)[13]Fine,Terrence L.[1999],Foundations of probability(update),in editors SamuelKotz,Campbell B.Read,David L.Banks Encyclopedia of Statistical Sciences, Update,3,John Wiley&Sons,NY.(volumes produced in the years1997–1999).[14]Fine,Terrence L.[2006],Probability and Probabilistic Reasoning for ElectricalEngineering,Pearson/Prentice Hall,Upper Saddle River,New Jersey. [15]Finetti,Bruno[1974],Theory of Probability,vols.I,II,translated by AntonioMachi and Adrian Smith,John Wiley&Sons,reprinted in1999.[16]Fitelson,Branden,Alan H´a jek,Ned Hall[2006],“Probability,”in JessicaPfeifer and Sahotra Sarkar,editors,The Philosophy of Science:An Ency-clopedia,Routledge Press,NY.[17]Hacking,Ian[1975],The Emergence of Probability,Cambridge UniversityPress,Cambridge.2[18]H´a jek,Alan[2003],“Interpretations of probability,”inE.N.Zalta,editor,The Stanford Encyclopedia of Philosophy,</archives/sum2003/entries/probability-interpret/>.[19]Jauch,Josef[1968],Foundations of Quantum Mechanics,Addison-Wesley Pub.Reading,MA.[20]Jeffrey,Richard[1992],Probability and the Art of Judgment,Cambridge Uni-versity Press,Cambridge.This is a collection of previously published articlestogether with updates and a new introductory essay.[21]Kaplan,Michael A.[1974],Extensions and Limits of Comparative ProbabilityOrders,Ph.D.dissertation,Cornell University,Ithaca.[22]Kaplan,Michael A.and T.L.Fine[1977],Joint orders in comparative proba-bility,Annals of Probability,5,161–179.[23]Keynes,John Maynard[1921],A Treatise on Probability,reprinted in1962byHarper Torchbooks,The Science Library,Harper&Row Pubs.,New York.[24]Kolmogorov,Andrei N.[1956],Foundations of Probability,translated byNathan Morrison,Chelsea Publishing Co.,NY.Translation of1933Grund-begriffe der Wahrscheinlichkeitsrechnung.[25]Krantz,David,R.Duncan Luce,Patrick Suppes,Amos Tversky[1971],Foun-dations of Measurement Vol.I:Additive and polynomial representations,Aca-demic Press,NY.[26]Kumar,Anurag[1982],Lower Probabilities on Infinite Spaces and Instabilityof Stationary Sequences,Ph.D.dissertation,Cornell University,Ithaca,NY.[27]Kyburg,Henry,E.jr.[1974],The Logical Foundations of Statistical Inference,D.Reidel Pub.,Dordrecht-Holland and Boston,MA.[28]Kyburg,Henry E.jr,Choh Man Tong[2001],Uncertain Inference,CambridgeUniversity Press.[29]Levi,Isaac[1980],The Enterprise of Knowledge:An essay on knowledge,credalprobability and chance,MIT Press,Cambridge,MA.[30]Li,Ming,Paul Vit´a nyi[1997],An Introduction to Kolmogorov Complexity,second edition,Springer-Verlag,New York.[31]Liboff,Richard[1992],Introductory Quantum Mechanics,2nd edition,Addison-Wesley Pub.,Reading,MA.[32]Luce,R.Duncan[1996],The ongoing dialog between empirical science andmeasurement theory,J.of Mathematical Psychology,40,78–98.[33]Mises,Richard von[1919],Grundlagen der Wahrscheinlichkeitsrechnung,Math.Zeit.,5,52–99.[34]Mises,Richard von[1957,1981],Probability,Statistics and Truth,second re-vised English edition,translation prepared by Hilda Geiringer,Dover Publi-cations Inc.,New York.Original edition in German in1928was published byJ.Springer.This is a translation of the authoritative third German edition asrevised by Mises and published in1951.[35]Mises,Richard von[1964],Mathematical Theory of Probability and Statistics,edited and complemented after his death by Hilda Geiringer(Mrs.von Mises),Academic Press,New York.[36]Narens,Louis[1985],Abstract Measurement Theory,MIT Press,Cambridge.[37]Neumann,John von[1947],The mathematician,Works of the Mind,I number1,University of Chicago Press,Chicago,180–196.3[38]Neumann,John von[1955],Mathematical Foundations of Quantum Mechanics,translated by Robert T.Beyer,Princeton University Press,Princeton,NJ. [39]Peirce,Charles Sanders[1892],The doctrine of necessity examined,TheMonist,2,321–337.[40]Peirce,Charles Sanders[1910],Notes on the doctrine of chances,reprintedin The Essential Peirce,Vol.1,edited by N.Houser and C.Kloesel,Indiana University Press,Bloomington,1992.[41]Rego,Leandro C.,Terrence Fine[2005],Estimation of chaotic probabilities,Fabio Cozman,Robert Nau,Teddy Seidenfeld,eds.,Proc.Fourth Int.Symp.on Imprecise Probabilities and Their Applications,SIPTA,Pittsburgh,297–305.[42]Rego,Leandro C.[2007],Conditioning in chaotic probabilities interpreted as ageneralized Markov chain,Gert de Cooman,Jirina Vejnarova,Marco Zaffalon, eds.,5th Inter.Symp.on Imprecise Probability:Theories and Applications, SIPTA,Prague,365–374.[43]Sadrolhefazi,Amir[1990],Interval-Valued Probability-based Time Series Mod-els with Measure-like marginals,Ph.D.dissertation,Cornell University,Ithaca, NY.[44]Sadrolhefazi,Amir,Terrence Fine[1994],Finite-dimensional distributions andtail behavior in stationary interval-valued probability models,Annals of Sta-tistics,22,1840–1870.[45]Shafer,Glenn,Vladimir Vovk[2006],The sources of Kolmogorov’s Grundbe-griffe,Statistical Science,21,70–98.[46]Solomonoff,Ray J.[1964],A formal theory of inductive inference,parts1and2,Information and Control,7,1–22,224–254.[47]Suppes,Patrick[1984],Probabilistic Metaphysics,Basil Blackwell,Oxford.[48]Suppes,Patrick[2002],Representation and Invariance of Scientific Structures,CSLI(Center for the Study of Language and Information)Stanford,CA. [49]Varadarajan,V.S.[1962],Probability in physics and a theorem on simultaneousobservability,Comm.on Pure and Applied Mathematics,X V,189–217. [50]Venn,John[1888],The Logic of Chance,third edition,reprinted in1962byChelsea Pub.,New York.[51]Walley,Peter,Terrence Fine[1982],Towards a frequentist theory of upper andlower probability,Annals of Statistics,10,741–761.[52]Walley,Peter[1991],Statistical Reasoning with Imprecise Probabilities,Chap-man and Hall,London.[53]Weinberg,Steven[1989],“The Cosmological Constant Problem,Review ofModern Physics,61,1–23.4。