《纳氏英文文法》第二册阅读趣话 共61页PPT资料
Lesson11ButWhatsaDictionaryFor课文讲解PPT讲解学习
• Dictionary:
• The first modern examples of lexicography are thought to be Nathan Bailey's Universal Etymological English Dictionary (1721) and his larger Dictionarium Britannicum (1730), which served Samuel Johnson, who was considered as England's first complete man of letters, in preparing his Dictionary of the English Language (1755), the first comprehensive English lexicography.
Summary and comments (16-17)
Part II (4-30) body argument
Return to thesis with new idea: Certainty impossible and simplification misleading (18)
Details with examples: spelling (19) pronu.(20-21)
Give facts or other usual data to support your thesis. This can be done in several paragraphs in order of importance or relevance to the thesis.
By tying together what has been discussed, conclude your essay with a restatement of the thesis or a forceful sentence that helps to drive your thesis home.
new2
Active reading 2: Navigation
Text organisation
A Marriage Valentine
to anyone to the person receiving the Valentine gift
He wishes for the cloths of heaven
More
Active reading 2: Destination
Summarising
The author of Valentine uses a dramatic monologue, a ____s_p_e_e__c_h_____ by a single ______c_h_a_r_a_c__te_r____ who shares thoughts, character and situation.
平生 以鸟儿的优雅姿态
已做完了一切， 此刻张开喙
丢下 一声叹息
轻如羽毛。
Back
Active reading 2: Navigation
VALENTINE
Listen to the MP3 and recite the poem.
Not a red rose or a satin heart. 不是一朵红玫瑰或一颗亮晶晶的心。
Questions and answers
Who is the poet speaking to in each poem? A marriage: to anyone, including perhaps himself, in a monologue. Valentine: to the person receiving the Valentine gift. He wishes for the cloths of heaven: to the person he loves.
paper illuminate intersec
Accident Analysis and Prevention 53 (2013) 65–77Contents lists available at SciVerse ScienceDirectAccident Analysis andPreventionj o u r n a l h o m e p a g e :w w w.e l s e v i e r.c o m /l o c a t e /a apTo illuminate or not to illuminate:Roadway lighting as it affects traffic safety at intersectionsJohn D.Bullough a ,Eric T.Donnell b ,Mark S.Rea a ,∗a Lighting Research Center,Rensselaer Polytechnic Institute,21Union Street,Troy,NY 12180,USAbDepartment of Civil and Environmental Engineering,The Pennsylvania State University,University Park,PA 16802,USAa r t i c l ei n f oArticle history:Received 14May 2012Received in revised form 3October 2012Accepted 20December 2012Keywords:Lighting SafetyRegression analysis Visual performance Crash reductionables to estimate the association between lighting and nighttime and daytime crashes and the resulting crash ratios.Overall,the presence of roadway intersection lighting was found to be asso-with an approximately 12%lower night-to-day crash ratio than unlighted intersections.In the analytical approach,visual performance analyses based on roadway intersection lighting prac-in Minnesota were made for the same intersection types investigated in the statistical approach.The results of both approaches were convergent,suggesting that visual performance improvements from roadway lighting could serve as input for predicting improvements in crash frequency.A provisional transfer function allows transportation engineers to evaluate alternative lighting systems in the design phase so selections based on expected benefits and costs can be made.© 2012 Elsevier Ltd. All rights reserved.1.IntroductionA primary purpose for installing roadway lighting is to increase the visual range afforded by vehicle headlamps while driving at night (IES,2000).It is generally believed that roadway lighting improves safety by reducing the frequency of crashes occurring at night through improvements in driver visibility.Some studies have examined statistical associations between roadway lighting presence and traffic safety,making no explicit assumptions about the role of visibility in mitigating nighttime crashes.Generally,the results from these studies (IES,1989;Elvik,1995)have led to the general conclusion that roadway lighting is associated with a reduc-tion in the night-to-day crash ratio (CIE,1992).A night-to-day crash ratio reduction of approximately 30%has been suggested for the overall crash safety effect from roadway lighting (CIE,1992).Most studies of roadway lighting and safety consider lighting as a binary variable (i.e.,present or not),but several attempts to relate specific characteristics of roadway lighting and safety have been made,using various photometric measures as surrogates for visibility.One early study was undertaken by Box (1971)where he compared the ratio of night-to-day crash rates along 22lighted and unlighted highway sections.The author grouped the highway∗Corresponding author.Tel.:+15186877100;fax:+15186877120.E-mail addresses:ream@ (M.S.Rea).URL: (M.S.Rea).sections into categories corresponding to mean horizontal illumi-nance levels between 3and 6lx,between 8and 11lx,and between 13and 15lx.Presumably,higher illuminances would result in greater visibility and might be expected to yield fewer nighttime crashes.The lighted sections had lower night-to-day crash rate ratios than the unlighted sections,but the lowest night-to-day crash rate ratio was found for the 3-to-6lx category,with higher night-to-day crash rate ratios for the two higher illuminance cat-egories.In a different study,Box (1976)evaluated the impact of reducing illuminance levels along a major highway from 14lx to 9lx and found that nighttime crash frequency increased by 10%,although daytime crash frequency also increased (by 4%).In com-parison,Scott (1980)measured roadway luminance levels along 89(each at least 1km in length)two-lane roadway sections.Despite a great deal of variability in night-to-day crash ratios among all of the sites,a best-fitting exponential function to the data yielded a monotonically decreasing relationship between night-to-day crash ratios and luminance level,consistent with Box (1976)but incon-sistent with Box (1971).The lack of agreement across these studies could be related to differences among the specific locations studied (e.g.,roadway geometry or traffic control)or different evaluation paradigms (e.g.,before/after or with/without comparisons).Even considering these differences,it is important to account for all the factors that could affect visibility.In particular,the impact of light-ing on visual performance depends upon the contrast and the size of a hazard,not simply upon illuminance or luminance levels provided by the roadway lighting system.0001-4575/$–see front matter © 2012 Elsevier Ltd. All rights reserved./10.1016/j.aap.2012.12.02966J.D.Bullough et al./Accident Analysis and Prevention53 (2013) 65–77In an attempt to relate specific measures of visibility associated with roadway lighting to safety,Janoff et al.(1978)studied night-time crash frequency and measured photometric conditions along several roadways,and found nighttime crash frequency to have a weak inverse relationship with a visibility metric they derived, which was based on the ratio between an object’s luminance con-trast and the contrast it would require to be just visible.Also,Keck (2001)summarized a different study of roadway lighting,visibil-ity and safety and reported that a similar visibility metric was not correlated with the frequency of nighttime crashes unless head-lamp illumination was also considered.Even then,the statistical association was modest in magnitude(coefficient of determination r2=0.12).In addition to the above issues related to study locations and evaluation methods,a factor that makes it difficult to assess the relationship betweenfixed roadway illumination systems and traf-fic safety is that lighting is installed for a variety of reasons(e.g., security,esthetics).Moreover,lighting is usually not the only road-way safety feature that is installed when a roadway is designed or improved(IES,1989).For example,lighting may be installed when converting a stop-controlled to a signalized intersection,or may be installed when threshold levels of pedestrian or vehicu-lar volumes,or annual crash frequencies,are exceeded.Thus,the estimates of association between roadway lighting presence and nighttime crashes in past studies may be confounded by other modifications to the study site.In the present paper,an exploratory strategy is presented to test the theoretical links among roadway lighting,visibility,and safety while accounting for as many potential safety-influencing variables as possible.To overcome previous limitations in the literature,our goal was to determine if we could establish convergence between statistical and analytical approaches to crash safety.Thus,both sta-tistical and analytical approaches were undertaken to relate crash frequencies to visual performance levels for the same set of roadway lighting conditions.Donnell et al.(2010)merged roadway lighting presence and roadway geometric and traffic volume and control data with daytime and nighttime crash data to assess the statis-tical association between roadway intersection lighting presence and the night-to-day crash ratio in Minnesota.In their study,many potential safety-influencing variables that were not considered in previous lighting-safety research were included in the statis-tical model estimation.In a parallel but independent investigative domain,Rea et al.(2010)developed photometric simulations of a large variety of roadway intersection lighting configurations to make context-specific predictions of visual performance levels pro-vided by roadway lighting systems varying in light level,spatial configuration,and ambient characteristics for drivers of different ages.Here we present a methodology to link roadway lighting char-acteristics to visual performance levels and,therefore,to traffic crashes.Importantly,our objective was to test whether there is convergence between statistical and analytical approaches,which would bolster the expected relationships among lighting presence, improved visibility,and improved safety.Each approach reduces the inherent uncertainty associated with the other by providing an independent basis to explore the theoretical relationship between roadway intersection lighting and reductions in nighttime crashes through improvements in driver visual performance.The present paper summarizes the parallel statistical and ana-lytical approaches,developed by Donnell et al.(2010)and by Rea et al.(2010),respectively,used to probe the theoretical relation-ship described above.First,the background,methods and results of each independent approach are provided;next,the evidence for the convergence of these approaches is described.As stated above,the approaches used in this paper are described in previ-ous publications(Donnell et al.,2010;Rea et al.,2010);the present paper describes how these approaches were adapted for different roadway intersection types in Minnesota.If the necessary links between roadway lighting,visual perfor-mance and traffic safety can in fact be forged through converging approaches,it could then be possible for traffic engineers to improve traffic safety by considering the visual performance levels and the costs of proposed roadway illumination systems through a provisional proposed transfer function relating visibility from light-ing to crash safety.2.Description of parallel approachesIn the present section,the statistical modeling and the visual performance modeling are described together with thefindings from these two approaches.2.1.Statistical modeling2.1.1.BackgroundA significant body of published literature consistently shows thatfixed roadway lighting improves intersection safety.Obser-vational before-after studies and with–without cross-sectional comparisons have been used to support thesefindings.The meas-ures used to evaluate the effects of roadway lighting differ across studies.With regard to observational before-after studies,several authors used reported nighttime crash rates(Walker and Roberts, 1976;Lipinski and Wortman,1978)as the safety performance measure.These studies found that roadway lighting was associ-ated with reduced nighttime crash rates of45–52%.Other authors (Walker and Roberts,1976;Schwab et al.,1982;Green et al., 2003;Isebrands et al.,2004,2006)reported13–49%reductions in reported nighttime crash frequencies attributed to lighting.Reduc-tions in reported night-to-day crash ratios of22–40%have been documented in other observational before-after studies related to roadway lighting(Lipinski and Wortman,1978;Preston and Schoenecker,1999;Isebrands et al.,2004,2006).With regard to with–without cross-sectional comparisons,the published literature also indicates that roadway lighting appears to be an effective safety countermeasure.The results of previous stud-ies have found that nighttime crash rates are25%lower at lighted intersections when compared to unlighted intersections(Preston and Schoenecker,1999);reported nighttime crash frequencies are 39%lower at lighted intersections(Schwab et al.,1982).Night-to-day crash ratios are31%lower at intersections with lighting when compared to intersections without lighting(Isebrands et al.,2004, 2006).More recently,an estimate of the safety effect of at-grade inter-section lighting was published by Harwood et al.(2007),using the results of a meta-analysis of published literature from Elvik and Vaa (2004).Harwood et al.(2007)estimated that an appropriate acci-dent modification factor for the presence of lighting is0.96,or a4% reduction in total crashes after the installation of roadway lighting (this corresponds to a larger percentage of nighttime crashes,since roadway lighting is only of benefit at night).This result is published in thefirst edition of the American Association of State Highway and Transportation Officials’Highway Safety Manual(AASHTO,2010).While many of the previous lighting-safety studies appear to offer similar results using a variety of measures(i.e.,nighttime crash rates,nighttime crash frequencies,night-to-day crash ratios), there are limitations to these studies.First,much of the pub-lished literature evaluates the safety effects of roadway lighting based on crash rates or crash frequencies using reported crash data.Crash rates assume a linear relationship between crashes and traffic volume,which may not necessarily be the case,partic-ularly at intersection locations(e.g.,Poch and Mannering,1996;J.D.Bullough et al./Accident Analysis and Prevention53 (2013) 65–7767Bauer and Harwood,1999;Washington et al.,2005).Reported crashes are rare events that vary randomly over time;thus,the use of expected crash frequencies in lighting-safety studies may be more meaningful.Secondly,many of the previous studies did not consider the relationship between intersection crashes and other safety-influencing features present at intersections(e.g.,intersec-tion form,intersection angle,posted speed limits,cross-section dimensions of approach legs,etc.).Statistical models of expected crash frequency should be conditioned on intersection features thought to influence crashes.While Hauer(2005)offered the empirical Bayes observational before-after study(which considers expected crash frequency) as a superior method to evaluate the safety effects of a coun-termeasure,such a method is difficult to apply in the case of roadway lighting because installation of a lighting system is often accompanied by other intersection improvements(e.g.,conver-sion of a stop-controlled to a signalized intersection or addition of a turn-lane).Thus,use of an observational before-after study is complicated by the inability to identify sites where only lighting was installed as a traffic safety countermeasure.A with–without comparison is thus a logical alternative evaluation method.Sev-eral studies have used this method(Schwab et al.,1982;Preston and Schoenecker,1999;Isebrands et al.,2004,2006)to assess the safety effects of intersection lighting.Past with–without lighting-safety studies used reported nighttime crash frequencies or rates, or night-to-day crash ratios,which are subject to the method-ological limitations identified above.The present study estimates statistical models of nighttime and daytime crash frequencies as a function of intersection features and traffic control.This approach enables comparisons of expected nighttime crash frequencies at intersections with and withoutfixed lighting,and also enables com-parisons of expected night-to-day crash ratios at locations with and without lighting.As noted previously,nighttime crash fre-quencies and night-to-day crash ratios are common performance metrics in past lighting research.With regard to night-to-day ratios, expected daytime crash frequencies are used as a form of con-trol to assess the association between crashes and the presence of lighting,which is consistent with past research(e.g.,CIE,1992; Griffith,1994).Because lighting is not turned on during the day-time,it is anticipated that roadway lighting will not be associated with daytime crashes.If this assumption holds,then the nighttime crash frequency estimates will be used to assess the safety effects of lighting.However,if there is an association between lighting and daytime crashes(e.g.,Lacy et al.,2004),this same association would be present during the nighttime,which is the basis for using night-to-day ratios as the performance measure to evaluate light-ing effectiveness.In this case,the night-to-day ratio will be used to assess the effects of lighting.2.1.2.Statistical modeling methodIn consideration of the aforementioned issues,the present anal-ysis is based upon a with–without comparison of expected crash frequencies,conditioned on several intersection safety-influencing features,as the method to assess the safety effects of intersec-tion lighting presence on motor vehicle crash frequency.Datafiles from Minnesota,which were acquired from the Federal Highway Administration’s(FHWA)Highway Safety Information System(HSIS), contained the intersection-level traffic volume,geometric design, and roadway lighting presence data needed to estimate cross-sectional models of intersection safety.Four years(1999–2002, inclusive)of crash and roadway inventory data were available for the analysis.A total of6464intersections,all of which are maintained by the Minnesota Department of Transportation,were included in the analysis.Due to some missing geometric or traffic control data at some intersections,only22,058sample observa-tions(out of a possible25,832)were used for statistical modeling.Descriptive statistics for the variables used in the statistical analysis are shown in Table1.A framework to estimate the safety effects of intersection light-ing presence using Minnesota HSIS data is described by Donnell et al.(2010).The authors’described the data structures,statistical modeling alternatives,possible endogeneity of lighting,and various analysis taxonomies.The entirety of the discussion is not repeated here;however,a brief review of various modeling alternatives is provided.Possible modeling methods include negative binomial regression to account for overdispersion in crash data(see Poch and Mannering,1996;Bauer and Harwood,1996;Washington et al., 2005for intersection examples).Generalized estimating equations (e.g.,Lord and Persaud,2000;Wang et al.,2006;Donnell et al., 2009b)have been used to account for temporal correlation result-ing from observing crash data annually.Random effects negative binomial(Shankar et al.,1997;Chin and Quddus,2003)and nega-tive multinomial(Ulfarsson and Shankar,2003)models have been used to address issues of spatial and temporal correlation in crash data.Most recently,models of crash frequency have employed random parameters models(e.g.,Anastasopoulos and Mannering, 2009)to account for the unobserved heterogeneity present across analysis locations.Lord and Mannering(2010)describe the advan-tages and disadvantages of using all of these methods to estimate statistical and econometric models of crash frequency.As described by Lord and Mannering(2010),the negative bino-mial regression model accounts for overdispersion in crash data and is a straightforward model to estimate.Thus,the negative binomial model seems to be a logical choice to model expected intersection crash frequencies in the present study.The general functional form of the model is as follows:ln i=ˇX i+ε(1) where i=expected number of crashes per year at intersection i;ˇ=vector of estimable regression parameters;X i=vector of geo-metric design,lighting presence,and traffic volume data;and εi=gamma-distributed error term.The mean–variance relationship for the negative binomial regression model is as follows:Var(y i)=E(y i)[1+˛E(y i)](2) where Var(y i)=variance of observed crashes y at intersection i;E(y i)=expected annual crash frequency at intersection i;and ˛=overdispersion parameter.The method of maximum likelihood is used to estimate the regression parameters in the negative binomial regression model. Ulfarsson and Shankar(2003)found that the negative binomial regression model outperformed the random-effects negative bino-mial regression model when indicator variables to account for serial correlation were included in the model.Thus,the present study applies a negative binomial regression model with indica-tor variables to account for temporal correlation resulting from the inclusion of yearly crash counts in the model.A random parameters count regression model is also estimated in the present study to account for unobserved heterogeneity across observations.As described by Lord and Mannering(2010), the regression parameters estimated using a random parameters model can be written as follows:ˇi=ˇ+ i(3) where i=randomly distributed error term(Anastasopoulos and Mannering,2009found the normal distribution with mean0and variance 2to provide bestfit in crash frequency models).A simulation-based maximum likelihood procedure was used in the present study to estimate the parameters.In this pro-cess,200Halton draws were used because recent research by68J.D.Bullough et al./Accident Analysis and Prevention53 (2013) 65–77Table1Variable definitions and descriptive statistics for Minnesota roadway intersection and crash data(Donnell et al.,2010).Continuous variables Min.Max.Mean Standard deviation Night crash frequency(per year)0280.3660.969Day crash frequency(per year)055 1.121 2.457Major road average daily traffic4077,43082849381Percent heavy vehicles on major road061.118.888 5.109Minor road average daily traffic177,43031645179Categorical variables Proportion in sampleArea type indicator(1=urban/suburban;0=rural)1:0.446 0:0.554Traffic control indicator (1=signal;0=stop-control)1:0.137 0:0.863Lighting indicator(1=present;0=not present)1:0.421 0:0.579Intersection type indicator (1=skew;0=cross or tee)1:0.110 0:0.890Speed indicator a(1=50mph or greater;0otherwise)1:0.673 0:0.327No access control indicator a(1=no access;0=partial access control)1:0.943 0:0.057Depressed median indicator a(1=depressed median;0=barrier or no median)1:0.116 0:0.884Paved left-shoulder indicator a(1=paved shoulder;0=unpaved or no shoulder)1:0.458 0:0.542Paved right-shoulder indicator a(1=paved shoulder;0=unpaved or no shoulder)1:0.510 0:0.490a Data were used for the major intersecting roadway only.Anastasopoulos and Mannering(2009)indicate that such a pro-cedure produces accurate parameter estimates and converges more quickly than random draws.For a complete review of ran-dom parameters count models,readers are referred to Greene (2008).2.1.3.Statistical modeling resultsThe results of the statistical analyses are shown in Table2. Daytime and nighttime total crash frequency models are shown based on the negative binomial regression models without yearly indicator variables,negative binomial regression model with yearly indicators,and the random parameters negative bino-mial regression model.Negative binomial regression models for different intersection type subsets(e.g.,rural signalized,rural stop-controlled,urban signalized,etc.)are included in Donnell et al. (2009a).All of the geometric design and traffic control variables have the same sign and many have similar magnitudes in the day-time and nighttime crash frequency models across all formulations, suggesting that the effects of these variables do not change considerably between the daytime and nighttime periods or across regression methods.A notable exception to this is the lighting presence indicator variable.As would be expected,the presence of roadway lighting is associated with fewer expected nighttime crashes than the baseline of no lighting present.In the daytime crash frequency models,the lighting presence indicator is0.048, 0.050,and0.078,respectively,across the three model formulations. This shows that the expected crash frequency at intersections with lighting is higher than the baseline condition of no lighting present by about5–8%,respectively,across the regression methods.The small magnitude of the lighting presence variable could suggest that more crashes are expected at intersections with lighting as a result of luminaires(i.e.,fixed poles)present along the roadside. Assuming that this statistical association is consistent both day and night,the night-to-day crash ratio was used to adjust the nighttime crash frequency estimate relative to the daytime estimate.The expected percent difference in the night-to-day crash ratio was computed using Eq.(4).100(N/D)w−(N/D)wo(N/D)wo−1(4)where N=expected number of nighttime crashes for all inter-sections included in the analysis;D=expected number of daytime crashes for all intersections included in the analysis; w=intersections with lighting;and wo=intersections without lighting.The percent difference in the night-to-day crash ratio for total intersection crashes in Minnesota,based on the negative binomial model in Table2,is computed as follows:100(exp(−0.079×1)/exp(0.048×1))w−(exp(−0.079×0)/exp(0.048×0))wo(exp(−0.079×0)/exp(0.048×0))wo−1=−11.8%(5)For all Minnesota intersections in the analysis,the expected per-cent difference in the night-to-day crash ratio is approximately12%lower at intersections with lighting when compared to inter-sections without lighting when applying the negative binomialregression model with no yearly indicator variables.When usingthe negative binomial regression model with yearly indicators andthe random parameters negative binomial regression models,theexpected percent difference in the night-to-day crash ratio is−11.9J.D.Bullough et al./Accident Analysis and Prevention53 (2013) 65–7769 Table2Daytime and nighttime crash frequency models for all intersection crashes in Minnesota.Variable Negative binomial(no year indicators)Negative binomial(with year indicators)Random parametersDaytime model Nighttimemodel Daytime model NighttimemodelDaytime model Nighttime modelCoeff. (SE b)Coeff.(SE b)Coeff.(SE b)Coeff.(SE b)Coeff.(SE b)Scale c(SE b)Coeff.(SE b)Scale c(SE b)Constant−6.533*(0.123)−6.886*(0.185)−6.560*(0.126)−6.908*(0.186)−6.927*(0.115)0.593*(0.008)−7.033*(0.176)0.228*(0.013)Log major road average daily traffic(veh/day)0.601*(0.012)0.572*(0.018)0.598*(0.012)0.569*(0.019)0.606*(0.011)0.037*(0.001)0.579*(0.018)0.006*(0.001)Log minor road average daily traffic(veh/day)0.160*(0.005)0.127*(0.008)0.161*(0.005)0.126*(0.008)0.177*(0.005)0.007*(0.001)0.131*(0.008)0.023*(0.002)Percent heavy vehicles on major road(%)−0.009*(0.002)−0.017*(0.004)−0.012*(0.002)−0.019*(0.004)−0.027*(0.002)0.055*(0.001)−0.034*(0.004)0.044*(0.002)Area type indicator (1=urban/suburban; 0=rural)−0.100*(0.020)−0.421*(0.032)−0.101*(0.021)−0.421*(0.032)−0.155*(0.021)0.116*(0.011)−0.455*(0.033)0.031(0.017)Traffic control indicator (1=signal control;0=stop-control)0.646*(0.032)0.713*(0.041)0.647*(0.032)0.714*(0.041)0.722*(0.022)0.001(0.014)0.774*(0.037)0.039(0.021)Lighting indicator (1=present;0=not present)0.048*(0.022)−0.079*(0.034)0.050*(0.023)−0.077*(0.035)0.078*(0.023)0.041*(0.010)−0.110*(0.036)0.292*(0.015)Intersection type indicator (1=skew;0=cross or tee)0.488*(0.020)0.486*(0.032)0.480*(0.021)0.482*(0.033)0.476*(0.023)0.045*(0.021)0.381*(0.037)0.425*(0.031)Speed indicator a(1=50mph or greater; 0=otherwise)−0.160*(0.017)−0.113*(0.027)−0.153*(0.018)−0.107*(0.027)−0.092*(0.017)0.082*(0.011)−0.097*(0.027)0.279*(0.016)Access control indicator a (1=no control;0=partial control)−0.042(0.043)−0.016(0.051)−0.048*(0.042)−0.018(0.051)−0.096*(0.028)0.379*(0.009)−0.160*(0.040)0.610*(0.014)Depressed median indicator a (1=depressed median;0=barrier or no median)0.085*(0.028)0.172*(0.042)0.089*(0.028)0.176*(0.042)0.164*(0.025)0.050*(0.018)0.208*(0.038)0.007(0.026)Paved left-shoulder indicator a(1=paved shoulder; 0=unpaved or no shoulder)−0.118*(0.035)−0.291*(0.053)−0.117*(0.036)−0.295*(0.053)−0.172*(0.030)0.143*(0.011)−0.288*(0.046)0.016(0.018)Paved right-shoulder indicator a(1=paved shoulder; 0=unpaved or no shoulder)0.081*(0.035)0.217*(0.052)0.080*(0.036)0.220*(0.052)0.127*(0.029)0.119*(0.011)0.239*(0.045)0.046(0.016)Year2002indicator (1=year2002;0=otherwise)−0.011(0.030)−0.012(0.042)Year2003indicator (1=year2003;0=otherwise)0.249*(0.026)0.225*(0.038)Year2004indicator (1=year2004;0=otherwise)0.021(0.030)0.027(0.041)Dispersion parameter(˛)0.950*(0.016)0.896*(0.028)0.948*(0.019)0.895*(0.032)3.643*(0.098)3.080*(0.173)Number of observations =22,058LL(constantonly)=−30,979LL(full model)=−27,087LL(constantonly)=−15,888LL(full model)=−15,168LL(constantonly)=−30,932LL(full model)=−27,033LL(constantonly)=−15,863LL(full model)=−15,144LL(constant only)=−54,438LL(full model)=−25,560LL(constant only)=−19,367LL(full model)=−14,737The shaded row corresponds to the model values for lighting.a Variable represents data available for major road approaches only.b Standard error.c Scale is the scale of the distribution for the random parameters as shown in Eq.(3). *Statistically significant at95%confidence level.。
《纳氏英文文法》第二册阅读趣话
《纳氏英文文法》第二册阅读趣话 （封 面） （书 名）
《纳氏英文文法》第二册阅读趣话
（目 录）
底）ቤተ መጻሕፍቲ ባይዱ
（封
《纳氏英文文法》第二册阅读趣话
二、研 读
《纳氏英文文法》第二册阅读趣话
相关信息：
※ 《纳氏英文文法》：简称《纳氏文法》，全套四册共有几十本之多。 ※ J. C. Nesfield：全名John Collinson Nesfield（纳斯菲尔德），英 国著名语法专家。 ※ 麦克米兰出版公司：Macmillan Publishers Ltd. ※ 出版时间：1895年第一版，之后曾再版不少于30次。 ※ 该书译者：伦敦大学 徐兆熊（英汉本出版于1921年），留英学者。 ※ １９２６年，浮石人赵灼出版《纳氏英文法讲义》，这是中国第一本 英文语法书，用中文和英文讲解英文语法，全书四大册，是当时全中 国各大学都采用的教材。——江门日报（广东省）
《纳氏英文文法》第二册阅读趣话
群益书社（出版社）介绍： 群益书社由陈子沛、陈子寿两兄弟于1902年在湖南长沙创立，1907 年在位于上海福州路的惠福里创办分社。当时的惠福里小弄堂里还 有一个经营萧条，但在中国出版史上留下浓重一笔的 亚东图书馆。 亚东图书馆在陈独秀——《新青年》主编的一手策划下创办，到
《纳氏英文文法》第二册阅读趣话
澳大利亚国家图书馆回函一（08-12-1）： We expect to respond by: 08/12/08 02:40 pm We will send our response to the following email: yu_wr2004@. We noticed that the question's last useful date to you is: 08/12/08 02:45 pm We will make every effort to respond to your question within one week. However, complex enquiries relating to manuscripts, music, oral history or pictures may take up to four weeks. Thank you for using our service. National Library of Australia
英语辅修-学生学习资料-metaphor 2
“a figure of speech in which a name, action, or description is applied to another to suggest a likeness between them”The Winston DictionaryStructure: The signified is the signifier. 本体是喻体。
隐喻是不用功能词(function words) 的。
•Money is a bottomless sea, in which honor, conscience, and truth may be drowned.•金钱是无底的海洋，荣誉、良心和真理都可以淹没在其中。
•Self-love is a cup without bottom; you might pour all the great lakes into it, and not fill up.•自私是只没底的杯子，即使你把五湖之水倒进去，也难灌满。
•Your case is a horse of another color. (Yours is quite a different case. )••隐喻的结构特征•喻体直陈式•College is a comma of a sentence of life. 大学就是人生长句中的一个逗号。
•The lecture was a Sahara Desert.•喻体半隐式•以动物做喻•From every window heads were craning for a view of it.•每个窗口都有人探出头来看。
=stretch their heads like a crane 以动物做喻•She is a peacock in everything but beauty. —0. Wilde.•He doesn’t have an idea of his own. He just parrots what other people say.•他没有自己的观点，只会鹦鹉学舌。
外教社2024新世纪新编英国文学史教程 PPT课件Unit 10
Questions for Critical Thinking
1. How do you comment on the role of literary writing in the practice of feminism? 2. Please name one middleclass character in the novels you have ever read, and tell how differently this middleclass character is represented, compared with those from other classes. 3. What are the major characteristics of the new woman? Which female character in English literature can be labelled as “the new woman” and why?
Important Features of Women Novelists in the 19th Century
An Aspiration for Gender Equality a) Women’s literary achievements in the 19th century established for them an extraordinary reputation in history. b) Female writers spared no efforts to appeal for more freedom and rights. An Active Involvement in Public Affairs a) More and more female novelists took literature as an instrument of enlightening the mass. b) Female writers even played a hidden role in boosting national identity.
brooks
L EGAL U NCERTAINTY,E CONOMICE FFICIENCY, AND THE P RELIMINARYI NJUNCTION D OCTRINERichard R.W. Brooks* and Warren F. Schwartz**I NTRODUCTION (382)I.A RTICULATED S TANDARDS FOR P RELIMINARY I NJUNCTIONS (388)A. Traditional Approaches (389)B. The Error-Minimizing Leubsdorf-Posner Formulation (390)C. Our Alternative Formulation (393)II.M ODEL AND A NALYSIS (394)A. The Leubsdorf-Posner Formulation of the Traditional BalancingRule (396)B. If Preliminary Injunctions Were Never Available (397)1. Incentives when compensation is provided for nonobligatoryperformance (398)2. Incentives when there is no compensation for nonobligatoryperformance (399)C. If Preliminary Injunctions Were Always Available (399)1. Incentives when compensation is provided for nonobligatoryperformance (400)2. Incentives when there is no compensation for nonobligatoryperformance (401)D. Implications: An Asymmetry in Substantive Law (401)III.W HICH R ULE S HOULD B E E MPLOYED? (402)A. Interim-Efficiency Rule (403)B. Preliminary Injunction Liability Rule (405)C. Error-Minimizing Rule in Practice (407)C ONCLUSION (409)*Associate Professor, Yale Law School.**Professor Emeritus of Law, Georgetown University Law Center. We thank Bruce Ackerman, Ian Ayres, Bill Eskridge, Owen Fiss, John Langbein, Eric Posner, Judith Resnik, Alan Sykes, and participants at Georgetown and Northwestern workshops. Sabrina Charles and Albert Wang provided outstanding research assistance.38158:381382 STANFORD LAW REVIEW [Vol.I NTRODUCTIONIn this Article, we consider preliminary injunctions from a radically different perspective than that articulated in judicial opinions and prior legal scholarship. By conventional accounts, when confronted with uncertain legal entitlements, courts should consider preliminary awards only if adequate compensatory remedies are unavailable. The trouble with this “compensatory”view is that it is unresponsive to the ex ante behavioral consequences of legal uncertainty. When rights are uncertain, parties appreciate the full benefits oftheir conduct, but they discount harm to others of this conduct by the likelihoodthat they possess a legal entitlement to so act. Hence, individual incentives to behave efficiently are distorted by uncertain legal entitlements. Preliminary injunctions correct this distortion by wielding a stick and providing a carrot fora defendant who would otherwise discount damages given some positive probability that she may not have to pay them. The powerful stick in this example is the in terrorem damages that defendant will be required to pay if an injunction is granted and she violates it. The carrot is the reimbursement of compliance costs if defendant prevails at the end of the litigation. These penalties and rewards come into play only if the plaintiff decides to pursue the injunction, which is to say that the preliminary injunction doctrine takes the conduct decision out of the hands of the biased defendant and places it in the hands of plaintiff who, by design, faces the proper marginal costs and benefitsof the decision. Interestingly, although courts do not claim that they are promoting efficient behavior when granting preliminary injunctions, that characterization represents a good account for much of what courts are doing.Preliminary injunctions are broadly used. Parties seek injunctions to enjoin patent, copyright, and trademark infringement,1 corporate mergers,2 breaches1.See, e.g., SunTrust Bank v. Houghton Mifflin Co., 268 F.3d 1257 (11th Cir. 2001) (vacating an order granting a preliminary injunction to the owners of copyright in the novelGone with the Wind,which had enjoined the publication and distribution of The Wind Done Gone); Ty, Inc. v. Jones Group, Inc., 237 F.3d 891 (7th Cir. 2001) (affirming grant of preliminary injunction in trademark infringement case relating to Ty’s Beanie Babies toys);CNB Fin. Corp. v. CNB Cmty. Bank, No. CIV.A.03-6945(PBT), 2004 WL 2434878 (E.D.Pa. Sept. 30, 2004) (enjoining defendant from trademark infringement); Best Cellars, Inc. v.Grape Finds at Dupont, Inc., 90 F. Supp. 2d 431 (S.D.N.Y. 2000) (enjoining defendants fromtrade dress infringement and copyright infringement); Progressive Games, Inc. v. Shuffle Master, Inc., 69 F. Supp. 2d 1276 (D. Nev. 1999) (granting preliminary injunction in gambling machine patent infringement case).2.See, e.g., Mony Group, Inc. v. Highfields Capital Mgmt., L.P., 368 F.3d 138 (2d Cir. 2004) (enjoining dissenting shareholders from mailing proxy cards); Bernard v. Air LinePilots Ass’n, Int’l, 873 F.2d 213 (9th Cir. 1989) (granting injunction to set aside labor agreement); United States v. UPM-Kymmene Oyj, 2003-2 Trade Cas. (CCH) ¶ 74,101 (N.D.Ill. 2003) (enjoining merger as violative of the Clayton Act, 15 U.S.C. § 12 (2000)); In rePure Res., Inc., S’holder Litig., 808 A.2d 421 (Del. Ch. 2002) (enjoining exchange offer pending alteration of terms); Solar Cells, Inc. v. True N. Partners, L.L.C., C.A. No. 19477,2002 LEXIS 38 (Del. Ch. Apr. 25, 2002) (enjoining merger).November 2005] THE PRELIMINARY INJUNCTION DOCTRINE 383 of contract,3 nuisances,4 marriages,5entertainment,6 and even manner of dress.7 In fact, almost any activity one can imagine is potentially subject to legal restraint through preliminary proceedings.8 However, remarkably little attention has been paid to whether these proceedings tend to promote or discourage desirable activities. The neglect of this issue among law and economics scholars is particularly difficult to explain.9 The doctrine specifying when a court will grant a preliminary injunction is cast in terms with obvious economic content. The preliminary injunction is only to be granted if plaintiff will suffer significant harm and stands to recover inadequate damages if she prevails at the conclusion of the case. Moreover, the right to a preliminary injunction depends in large part (under all versions of the controlling rule) on3.See, e.g.,Arch Pers. Care Prods., L.P. v. Malmstrom, 90 Fed. Appx. 17 (3d Cir. 2003) (affirming order enforcing a noncompetition agreement); Northwest Bakery Distribs., Inc. v. George Weston Bakeries Distribution, Inc., No. 04-C8233, 2005 U.S. Dist. LEXIS 385 (N.D. Ill. Jan. 11, 2005) (allowing injunction to stop termination of bakery distribution agreement); Hillard v. Guidant Corp., 37 F. Supp. 2d 379 (M.D. Pa. 1999) (enjoining defendant from breaking exclusive sales contract); V.I. Taxi Ass’n v. V.I. Port Auth., 36 V.I.43 (1997) (enjoining defendants from violating a taxi-franchise agreement).4.See, e.g., City of S. Pasadena v. Slater, 56 F. Supp. 2d 1106 (C.D. Cal. 1999) (enjoining freeway extension); Bragg v. Robertson, 54 F. Supp. 2d 635 (S.D. W. Va. 1999) (enjoining issuance of mining permits); United States v. Power Eng’g Co., 10 F. Supp. 2d 1145 (D. Colo. 1998) (directing defendant to comply with state regulations for hazardous-waste facilities); Maloof v. State Dep’t of Env’t, 136 Md. App. 682 (Ct. Spec. App. 2001) (affirming circuit court’s issuance of preliminary injunction enjoining operation of landfill).5.See, e.g., Largess v. Supreme Judicial Court, 317 F. Supp. 2d 77 (D. Mass. 2004).6.See, e.g., Elvis Presley Enters. v. Passport Video, 349 F.3d 622 (9th Cir. 2003) (enjoining use of video clips); Video Pipeline, Inc. v. Buena Vista Home Entm’t, Inc., 342 F.3d 191 (3d Cir. 2003) (enjoining use of movie clip previews on the Internet); A&M Records, Inc. v. Napster, Inc., 239 F.3d 1004 (9th Cir. 2001) (challenging sharing and copying of music MP3 files); ABKCO Music, Inc. v. Stellar Records, Inc., 96 F.3d 60 (2d Cir. 1996) (enjoining copyright infringement by karaoke company); EMI Latin v. Bautista, No. 03 Civ. 0947 (WHP), 2003 U.S. Dist. LEXIS 2612 (S.D.N.Y. Feb. 24, 2003) (enjoining defendant from interfering with plaintiff’s rights to manufacture and distribute music album).7.See, e.g., Newsom v. Albemarle County Sch. Bd., 354 F.3d 249 (4th Cir. 2003) (using injunction to prevent the enforcement of a high school dress code); Luckette v. Lewis, 883 F. Supp. 471 (D. Ariz. 1995) (seeking preliminary injunction against a prison policy that prohibited the plaintiff from wearing colors thought to be associated with particular gangs).8.Preliminary injunctions have recently been sought in other areas, including preventing executions. In Germany v. United States, 526 U.S. 111 (1999), Germany sought a preliminary injunction to prevent the scheduled execution of one of its citizens in Arizona. See also Ozmint v. Hill, 541 U.S. 929 (2004) (granting application to vacate preliminary injunction of execution). Preliminary injunctions have also been used to prevent religious celebrations, Chabad of S. Ohio v. City of Cincinnati, 363 F.3d 427 (6th Cir. 2004), and to challenge courthouse and classroom postings of the Ten Commandments. ACLU of Ky. v. McCreary County, 354 F.3d 438 (6th Cir. 2003), aff’d, 125 S. Ct. 2722 (2005).9.There have been some efficiency considerations of preliminary injunctions in the context of intellectual property cases, but these works have not explored the efficiency of the standard per se. See, e.g., Jean O. Lanjouw & Josh Lerner, Tilting the Table? The Use of Preliminary Injunctions, 44 J.L.&E CON. 573 (2001); Mark A. Lemley & Eugene Volokh, Freedom of Speech and Injunctions in Intellectual Property Cases, 48 D UKE L.J. 147 (1998).58:381384 STANFORD LAW REVIEW [Vol. the probability that plaintiff will prevail if the case is litigated to a conclusion. These requirements seem, unmistakably, to represent an attempt to adapt efficiently to the uncertainty of the final outcome. In commonsense terms, thereis a prevailing awareness that the requisite tasks to achieve the objectives of the controlling legal rule cannot be deferred until the conclusion of the litigation.Preliminary injunction doctrine recognizes that the task of protecting legal entitlements cannot be postponed until the conclusion of the litigation concerning the assignment of those entitlements. This fact is at odds with the usual law and economics understanding that the assignment and protection of entitlements can be separated and handled sequentially so long as damages atthe conclusion of the case are adequate.10 Accordingly, proponents of the current preliminary injunction doctrine cite the oft-mentioned claim that adequate damages at the conclusion of the case make the entitlement holder whole while encouraging efficient allocation of resources. The most prominent expression of this claim is the so-called “efficient breach hypothesis.”11 This10.Building on Guido Calabresi and A. Douglas Melamed’s framework for analyzinghow legal entitlements are assigned and protected, scholars often look at entitlement protectionin isolation from the assignment of rights. Guido Calabresi & A. Douglas Melamed, Property Rules, Liability Rules, and Inalienability: One View of the Cathedral, 85 H ARV.L.R EV. 1089 (1972). Theories of efficient liability rules begin with certainty over some initial entitlement: Thecourt has already determined to whom the entitlement belongs (or the parties know how thecourt will rule). The parties also know the remedies. Armed with this information, efficient tradeoffs can be made. Unfortunately, parties frequently do not know to whom the court willgrant an entitlement. Preliminary injunctive actions make this indeterminacy quite apparent; yet,the indeterminacy is present in numerous other contexts.11.The efficient breach hypothesis has its origins in seventeenth-century British common law. See Bromage v. Genning,(1616) 81 Eng. Rep. 540 (K.B.). Oliver Wendell Holmes cites Lord Coke in Bromage for support of the fundamental premise of the efficient breach hypothesis: “The duty to keep a contract at common law means . . . you must pay damages if you do not keep it,—and nothing else.” Oliver Wendell Holmes, The Path of theLaw, 10 H ARV.L.R EV. 457, 462 (1897). Departing from this premise, the Holmesian “bad man”—knowing the costs and the benefits of contract completion—is in a good position to determine the efficient remedy. It appears that the term “efficient breach” may have been coined, at least in print, by Charles J. Goetz and Robert E. Scott. See Charles J. Goetz & Robert E. Scott, Liquidated Damages, Penalties and the Just Compensation Principle: SomeNotes on an Enforcement Model and a Theory of Efficient Breach, 77 C OLUM.L.R EV. 554 (1977). Earlier formal treatments of the efficient breach hypothesis were offered by Robert Birmingham, John Barton, and Richard Posner. See R ICHARD A.P OSNER,E CONOMICA NALYSIS OF L AW § 3.8 (1972); John H. Barton, The Economic Basis of Damages for Breach of Contract, 1 J.L EGAL S TUD. 277 (1972); Robert L. Birmingham, Breach of Contract, Damage Measures, and Economic Efficiency, 24 R UTGERS L.R EV. 273 (1970); Robert L. Birmingham, Damage Measures and Economic Rationality: The Geometry of Contract Law, 1969 D UKE L.J. 49. Calabresi and Melamed generalized the insight beyond contracts using liability rules, and this work was expanded and further formalized by Louis Kaplow and Steven Shavell. See Calabresi & Melamed, supra note 10; Louis Kaplow & Steven Shavell, Property Rules Versus Liability Rules: An Economic Analysis, 109 H ARV.L.R EV.713, 725 (1996) (arguing that liability rules (with appropriate court-determined damages) allow infringers, who know whether their own valuations exceed the court’s damages, to make an allocationally efficient choice from their more informed perspective).November 2005] THE PRELIMINARY INJUNCTION DOCTRINE 385 hypothesis maintains that court-ordered expectation damages (a liability rule) lead parties to maintain or abandon prior agreements efficiently. Although this argument was initially focused on contracts, similar efficiency-based arguments have also been made to promote the use of liability rules within the context of tort, property, corporate, and constitutional law.12The basic idea is that if a party is required to compensate anyone harmed by particular conduct, the party, in deciding whether, and with what frequency, to engage in the conduct, will internalize the costs imposed on others and engage in the conduct only to the point at which the benefits of doing so exceed the aggregate costs. Liability rules encourage parties to weigh the costs of avoiding liability—through performance (e.g., completing a contract) or nonperformance (e.g., not causing a nuisance or otherwise interfering with another’s entitlement)—against the costs of facing liability (e.g., breaching the contract and paying the damage remedy or causing a nuisance and paying compensation). Thus, when properly employed, the liability rule remedy, we are constantly reminded, maximizes social welfare. The starting point of our analysis is that such efficiency claims often are wrong.When the assignment of entitlements (and, hence, liability for interference with entitlements) is uncertain, parties rationally discount harms when selecting their course of conduct. Uncertainty biases the estimates that are required under the efficient breach and other efficient “takings” hypotheses. This kind of uncertainty is virtually always present in preliminary proceedings. Yet leading commentators see “no occasion to grant immediate protection” when a “final judgment can remedy the plaintiff’s injuries.”13 We show, however, that the availability of an adequate final remedy is not a sufficient justification for denying preliminary injunctions: adequate compensation at the conclusion of the case does not provide parties with sufficient incentive to engage in efficient conduct before and during the case.14 This point has been obscured by the static Thus, liability rules are able to harness the private information held by the relatively more informed infringers. But cf.Ian Ayres & Paul M. Goldbart, A Critique of “Tangibility” as the Basis of Probability Rules(Yale Law School, Program for Studies in Law, Economics, and Public Policy, Working Paper No. 251, 2002).12.See Ashutosh Bhagwat, Hard Cases and the (D)Evolution of Constitutional Doctrine, 30 C ONN.L.R EV. 961, 1008 (1998) (arguing that liability rules can protect constitutional rights more effectively than property rules in some cases); Calabresi & Melamed, supra note 10; Kaplow & Shavell, supra note 11; Eugene Konotorovich, Liability Rules for Constitutional Rights: The Case of Mass Detentions, 56 S TAN.L.R EV. 755 (2004). But cf. A KHIL R EED A MAR,T HE C ONSTITUTION AND C RIMINAL P ROCEDURE:F IRST P RINCIPLES 115(1997);Jules L. Coleman & Jody Kraus, Rethinking the Theory of Legal Rights, 95 Y ALE L.J. 1335, 1339-40 (1996).13.John Leubsdorf, The Standard for Preliminary Injunctions, 91 H ARV.L.R EV. 525, 541 (1978). Even efficiency-minded judges have echoed this view. See Am. Hosp. Supply Corp. v. Hosp. Prods., 780 F.2d 589, 594 (7th Cir. 1986).mentators have emphasized the incommensurability of damages for certain temporary losses of entitlements: “The right to speak or vote or worship after trial does not replace the right to speak or vote or worship pending trial, and damages for temporary loss58:381386 STANFORD LAW REVIEW [Vol. treatment of preliminary injunctions in the existing academic literature.15 However, it is clear that the preliminary injunctive proceeding is a dynamic process—a process that compels consideration of ex ante motivations and strategic behaviors.16 Viewed from this light, preliminary injunction doctrinecan clearly be seen as an adaptive response to the impairment of parties’ incentives resulting from the uncertainty of entitlement assignments.For concreteness, consider the following hypothetical involving a contractfor the provision of a well-specified good by a seller to a buyer who has paid afixed amount up front.17 If we set the seller’s cost to 70 and the buyer’s valueto 100, performance of the contract would increase social welfare by placingthe good in the hands of the higher-valuing party (the buyer, in this case). The possibility of expectation damages makes it in the personal interest of the sellerto do what is socially desirable. If she does not perform, the seller saves 70 in terms of performance costs but must pay 100 in the form of expectation damages to the buyer. The remedy thus aligns the seller’s incentives with that which is socially desirable. However, this simple implication does not hold when liability is uncertain—a state of the world which, we again emphasize, is reasonably presumed in the context of preliminary injunction hearings.Uncertainty over entitlements changes the efficient breach calculation. For example, imagine that the seller believes there is a 50% chance that her obligation to perform, under the prevailing circumstances, will be legally excused. Under these circumstances, she will not perform (though performance results in the most efficient result), even when expectation damages are perfectly estimated and fully compensatory. When deciding whether to perform, the seller still compares the expected cost of performance (70) to the expected damages for breaching. In this case, however, her expected damagesare now 50, reflecting the expectation damages (100) discounted by the likelihood that the seller will not be held liable for breach (50%).18 So the rational, risk-neutral seller will not perform even though she may be required to make the buyer whole ex post. This is not an efficient result. Liability rules generally (and expectation damages specifically) do not preserve parties’of such rights are not even approximate compensation.” D OUGLAS L AYCOCK,T HE D EATH OFTHE I RREPARABLE I NJURY R ULE 122 (1991). While this point has merit, it is not the one we advance here. Our claim is that even when damages provide approximate and adequate compensation for individual harm, avoidable efficiency losses still accrue.15.Cf. Lanjouw & Lerner, supra note 9.16.Strategic use of preliminary injunctions by plaintiffs is not uncommon. Partiesoften pursue preliminary actions, knowing that they are likely to get the same judge at thefinal stage (especially in state courts) and that judge is unlikely to switch her views of themerits subsequently. This may improve a party’s bargaining power in settlement negotiationsor may offer some other strategic advantage over competitors. See id.17.The upfront payment simplifies our example by allowing us to focus exclusively onthe seller’s breach decision.18.The expected cost of breaching is now a 50% chance of owing 100 and a 50% chance of owing 0 (i.e., (100 × 50%) + (0 × 50%) = 50).November 2005] THE PRELIMINARY INJUNCTION DOCTRINE 387 incentives to behave efficiently in the context of legal uncertainty19—the quintessential context of preliminary hearings.20The preliminary injunction restores efficiency to liability rules by taking the breach decision out of the hands of the compromised seller. At the point where the seller announces that she is going to breach, the buyer (if she does nothing) expects a 50% chance of receiving 100 and nothing otherwise, leaving her with an expected value of 50. If, however, the buyer seeks a preliminary injunction, she will receive a value of 100 through performance, which represents a net expected increase in value of 50 (i.e., 50% × 100) at an expected cost of only 35 (i.e., 50% × 70, representing the 50% chance that she will have to reimburse the seller’s compliance cost if her injunction was granted improperly). As a general matter, it is easy to show that the buyer will compel performance through a preliminary injunction if, and only if, performance is efficient.21 Thus in the presence of legal uncertainty, a key (but largely unappreciated) function of preliminary injunctions is to promote efficiency.Of course, indeterminacy of entitlement is not limited to cases involving preliminary injunctions; it is not uncommon for parties to be unsure of their future liabilities at the stage at which the efficient breach (or taking) hypothesis requires them to calculate the expected costs and benefits of their planned conduct. In these situations, in which the rights to be determined by the proceedings are uncertain, the promised efficiency of liability rules cannot be assured. Therefore, the implications of our analysis reach beyond preliminary injunctions. As a point of departure into our analysis, we now briefly consider the various legal rules which are, or might be, used in deciding whether to granta preliminary injunction.19.Often, with liability rules, defendant-infringer is well positioned to weigh the costs and benefits of her conduct. But uncertainty over entitlements biases her view of the benefits because she is liable for compensating unrealized benefits in only some cases when she infringes, while she faces the costs in all cases when she does not infringe. Preliminary injunctions respond to infringers’ discounting of damages by providing compensation to defendants who are compelled to engage in conduct that is later determined not to be legally required.20.It is also important to emphasize that our results do not rest on an assumption of symmetric perfect information among the parties. In fact, the most significant implication of our model comes from an asymmetric information framework. That is, awarding a preliminary injunction in essence converts a property rule into a temporary liability rule: the party seeking the injunction has a limited call option, wherein she will have to pay court-determined damages if it turns out that the entitlement belongs to the other party. Knowing her private value and the distribution of harms to the other party from complying with the injunction, the liability-rule-like preliminary injunction harnesses the party’s private information. See, e.g., Kaplow & Shavell, supra note 11.21.More generally, if the probability that the court will ultimately award the entitlement to defendant is equal to (1 – P), lπ is plaintiff’s value, and l∆ is defendant’s value, then plaintiff will only seek a preliminary injunction if (1 – P) × lπ > (1 – P) × l∆. See discussion in Part III for a more formal and complete statement of this inequality.388 STANFORD LAW REVIEW [Vol.58:381I.A RTICULATED S TANDARDS FOR P RELIMINARY I NJUNCTIONSAs parties compete to prohibit or permit legally uncertain activities, courtsare asked to allocate consequential legal entitlements in preliminary proceedings without the benefit of a full hearing. In his now-classic study, The Standard for Preliminary Injunctions, John Leubsdorf describes numerous, often inconsistent, articulated bases for preliminary relief: “Irreparable injurymay or may not be mentioned. Sometimes the injunction must not disserve the public interest, sometimes it must serve the public interest, and sometimes onlythe equities of the parties count.”22 Sometimes the decision turns on maintaining the status quo, and other times facilitating change is key.23 Articulated constraints on the merits of plaintiffs’ claims indicate a similar lackof consistency: “One line of cases requires plaintiffs to show a fair question onthe merits, another a substantial probability of success, another a reasonable certainty, and another a clear right.”24Professor Leubsdorf suggests a coherent rationale that underlies this apparent confusion in the judicial opinions. The objective, according to this rationale, is to prevent irreparable injury to the parties’ legal rights.25 Historically, courts of chancery issued injunctions to prevent actions at law aswell as to preserve them. Leubsdorf argues that this latter concern (i.e., preserving and protecting rights at law) was the precursor to contemporary preliminary injunction analysis.26 From this historical line, it is not difficult tosee how one could arrive at a standard based on the prevention of irreparable injury to legal rights.Yet, equitable courts sometimes issued preliminary injunctions that werenot contingent on actions at law or actions in other courts. “In some instances, plaintiffs in suits properly instituted in Chancery needed immediate relief pending the Chancellor’s decision on the merits. They might seek, for example,to stop equitable waste or to secure interim enforcement of a contract.”27 Departing from this historical line of cases might lead one to emphasize a different standard for preliminary relief—a standard, for example, based on the avoidance of waste or efficient enforcement of contracts.2822.Leubsdorf, supra note 13, at 526 (footnotes omitted).23.See Thomas R. Lee, Preliminary Injunctions and the Status Quo, 58 W ASH.&L EEL.R EV. 109, 111 n.4 (2001).24.Leubsdorf, supra note 13, at 526 (footnotes omitted).25.Clearly visible traces of this objective were observable in eighteenth-century English common law courts, where the separation of legal and equitable proceedings frequently prompted judges to issue preliminary relief to protect rights at equity and law. Id.at 527-32.26.Id. at 528 (“Only [this class of cases] raised problems calling for the kind of analysis we now apply, and the particularities of that class have shaped later learning on the standard for preliminary injunctions.”).27.Id. at 529.28.Indeed, as Professor Leubsdorf notes, “[o]wing in part to Jeremy Bentham . . . . theNovember 2005] THE PRELIMINARY INJUNCTION DOCTRINE 389 In his essay, Leubsdorf himself was quite hostile to the notion of judges making utilitarian calculations when considering the issuance of preliminary injunctions:29 “The court’s function is to protect rights, not to increase the gross national product.”30 However, when rights are uncertain, as they are in preliminary hearings, why shouldn’t courts consider whether granting the injunction will increase social welfare or avoid waste? This, admittedly, is not among the stated justifications for preliminary injunctions advanced in judicial opinions. These opinions, for the most part, take an ex post view, in which the articulated concern is that damages awarded at the conclusion of the case may be inadequate to compensate plaintiff for harm suffered during the pendency of the case.31A. Traditional ApproachesDespite the rhetorical variation in the case law, there is a widely shared view that the purposes served by preliminary injunctions are maintaining the status quo between the parties, preserving the court’s ability to consider the case fully, and minimizing the harm caused by erroneous preliminary decisions. There is less apparent consensus on the best way to achieve these ends through the use of preliminary injunctions. Most courts, when deciding whether to grant an injunction, rely on a four-part standard that (to varying degrees) considers (1) plaintiff’s likelihood of success on the merits, (2) the amount of irreparable utilitarian tide [that swept] away the division between law and equity [also] shaped thought on preliminary injunctions, setting that thought adrift in search of a new foundation.” Id. at 532 (footnotes omitted). Could this tide not have carried judicial thought about preliminary injunctions to utilitarian shores? After all, it was around this time that “courts came to require plaintiffs seeking interlocutory relief to accept liability for resulting damage” to defendants. Id. at 534. And this requirement is, as we argue below, an essential move to assure efficiency through the use of preliminary injunctions.29.Leubsdorf maintained this attitude: “Even in common law nuisance cases where the court must perform something like a cost-benefit analysis at trial, other considerations control at the interlocutory hearing.” Id. at 543 n.101.Leubsdorf felt that at the interlocutory hearing, the court’s “goal is to assess the probable loss of rights under various courses of action, not to appraise the net social benefit from those courses.” Id. (emphasis added).30.Id. at 555. Efficiency could never be a sufficient warrant for “interim accommodation,” according to Leubsdorf. Irreparability is always a necessary condition: “The plaintiff must always demonstrate the possibility of irreparable loss.” Id. at 551. Leubsdorf did, however, recognize one category of preliminary injunctions—so-called “statutory injunctions”—in which social policy may justify preliminary intervention without a showing of irreparability. In these instances, “[t]he goal is not to minimize loss of rights in specific cases, but to isolate classes of cases in which granting or denying relief under specified tests will minimize harm to public policies.” Id. at 565. But, of course, public policy might reasonably include increasing aggregate social welfare or the gross national product.31.See, e.g., Am. Hosp. Supply Corp. v. Hosp. Prods., 780 F.2d 589, 594 (7th Cir. 1986) (“The premise of the preliminary injunction is that the remedy available at the end of trial will not make the plaintiff whole; and, in a sense, the more limited that remedy, the stronger the argument for a preliminary injunction . . . .”).。
LANGUAGES
U LRICH F RANK T OWARDS A S TANDARDIZATION OF O BJECT-O RIENTED M ODELLINGL ANGUAGES?Mai 1997Arbeitsberichte des Instituts für Wirtschaftsinformatik Nr. 3U LRICH F RANK T OWARDS A S TANDARDIZATION OF O BJECT-O RIENTED M ODELLINGL ANGUAGES?Mai 1997Arbeitsberichte des Instituts für Wirtschaftsinformatik Nr. 3Die Arbeitsberichte des Instituts für Wirtschaftsinfor-matik dienen der Darstellung vorläufiger Ergebnisse, die i.d.R. noch für spätere Veröffentlichungen überar-beitet werden. Die Autoren sind deshalb für kritische Hinweise dankbar.The "Arbeitsberichte des Instituts für Wirtschaftsin-formatik" comprise preliminary results which will usually be revised for subsequent publications. Criti-cal comments would be appreciated by the authors.Alle Rechte vorbehalten. Insbesondere die der Über-setzung, des Nachdruckes, des V ortrags, der Entnah-me von Abbildungen und Tabellen - auch bei nur auszugsweiser Verwertung.All rights reserved. No part of this report may be re-produced by any means, or translated.Arbeitsberichte des Instituts für WirtschaftsinformatikHerausgegeben von / Edited by:Prof. Dr. Ulrich FrankProf. Dr. J. Felix Hampe Prof. Dr. Stefan KleinBezugsquelle / Source of Supply:Institut für Wirtschaftsinformatik Universität Koblenz-Landau Rheinau 156075 Koblenz Tel.:0261-9119-480Fax:0261-9119-487Email:iwi@uni-koblenz.deWWW:http://www.uni-koblenz.de/~iwiAnschrift der Verfasser/Address of the authors:Prof. Dr. Ulrich FrankInstitut für Wirtschaftsinformatik Universität Koblenz-Landau Rheinau 1D-56075 Koblenz©IWI 1997AbstractObject-oriented modelling is used in a growing number of commercial software development projects. However, the plethora of approaches and corresponding CASE tools still prevents corporate users to migrate to object-oriented software development methods. Against this background the recent efforts of the Object Management Group (OMG) to standardize object-oriented modelling languages seem to promise substantial benefits: Not only will a standard allow to easily port a model from one CASE tool to another, it will also protect investment in training. In addition, it is a prerequisite for standardized business object models which - in the long run - may substantially improve the economics of developing and maintaining corporate information systems. Nevertheless there are objections against a standardization at present time: It is questionable, whether the state of the art in object-oriented modelling is mature enough to allow for standardization. Furthermore standardization holds the risk to discourage further innovations.In order to analyze the state of the art, this report will first give an overview of previous ap-proaches to object-oriented modelling. To get an idea of the present state of the art, the essen-tial characteristics of the modelling languages currently under review at the OMG are briefly characterized - together with an additional approach that recently has gained remarkable atten-tion. This will lead to a number of research challenges still to overcome. Finally the report of-fers a subtly differentiated answer to the question whether or not it is time for standardizing object-oriented modelling languages.1.MotivationDuring the last decade, object-oriented software development has been adopted in the academ-ic world with remarkable enthusiasm. This is different with corporate software development: Although there is a tremendous marketing push caused by vendors, consultants, and specialist journals, many companies are hesitating to introduce object-oriented methods1. This is for comprehensible reasons: At present time, there are still serious inhibitors of object-oriented software development to overcome. There is lack of mature technology. While there are ob-ject-oriented programming languages that come with reliable compilers - sometimes embed-ded in rather convenient environments, today’s object-oriented database management systems are usually not suited to replace relational database management systems within corporate in-formation systems. Furthermore there is lack of competence. In order to exploit the benefits offered by the object-oriented paradigm, and to avoid its pitfalls, it is necessary for the devel-opers to gain a deep understanding of the essential concepts. It is certainly not too daring to assume that most software developers do not have these skills today. From a managerial point of view it is risky to provide for the training that is required to develop these skills: Not only that training is expensive, and its success is hard to predict, furthermore there is still lack of standards. This is the case for object-oriented programming languages and database manage-ment systems, as well as for specialized development methods. For this reason protection of investments, both in professional training and in technology (CASE, compilers, etc.), is usual-ly not satisfactory.Although many companies are still hesitating to introduce object-oriented methods and tech-nologies, there is no doubt that the future (that is at least the next ten years) of corporate soft-ware development will be more and more object-oriented - simply because there is no alterna-tive paradigm of similar relevance. In other words: The already big market for object-oriented concepts, training, and technologies can still be expected to grow at a fast pace. However, in order to encourage corporate investments, it is not sufficient to develop mature technologies. In addition it is crucial to provide reliable standards that foster interoperability and reusability, and protect investments at the same time. Only recently it became apparent that the various di-alects of object-oriented modelling will eventually be replaced by an industry standard mod-elling language. On the one hand, three of the best known authors of modelling methods de-cided to merge their efforts into one company in order to consolidate their methods into one "unified method" ([Rum96b]). On the other hand, the "Object Management Group" (OMG) is-sued a request for proposals for object analysis and design ([OMG96]). The submissions were due at January, 17th, 1997. The review process can be expected not to be completed before the end of 1997. The OMG’s request for proposal was the cause for writing this report which will focus on the following questions:•What is/could be subject of standardization in the area of object-oriented modelling?•What are the prospects and risks that are related to standardization?•What is the OMG’s intention?•What are the characteristics of the official proposals to the OMG?1. To our knowledge there has not been a single representative study on an international scale. A recentlyconducted empirical study ([Sch97]) indicates that less than 10% of the german insurance companiesuse object-oriented software development methods.•Is the current state of the art mature enough to allow for standardization?This paper is intended to be followed by a later report ([FrPr97]) which will compare two mod-elling languages, the UML ([Rat97a] .. [Rat97j]) that has been proposed to the OMG, and the OML [FiHe96].2.Object-Oriented Modelling: Need for Standardization?During the last 10 years a remarkable number of approaches to support object-oriented soft-ware development on a conceptual level have been published. The majority of them evolved in an academic setting. Unfortunately, most of them are not documented in a comprehensive way. Furthermore the differences between some approaches seem to be marginal. For this rea-son, it would not make much sense to consider all of the approaches we have encountered. In-stead we will give an overview that is restricted to a few characteristics. With the increasing popularity of object-oriented software development, object-oriented modelling has gained more and more commercial attention - resulting in a growing market for specialized services and tools. Against this background - a new paradigm, attractive both from an academic and a commercial point of view - it is not surprising that there is remarkable diversity in concepts and terminology. In order to discuss the benefits and risks of standardization we will define essential subjects of object-oriented modelling - such as method, model, metamodel, etc. 2.1Overview of Previous Object-Oriented Modelling MethodsObject-oriented software development in general, object-oriented modelling in particular has been a rather popular topic both for conferences and specialized journals. This may indicate a vivid period of research - or simply the usual hyper-activity caused by the emergence of a new research paradigm (or at least what seems to be a new paradigm). To give an impression of the amount of research, we have gathered more than 40 approaches which focus on various aspects of object-oriented analysis and design. The following list is compiled from a previous survey ([Fra93]) and additional sources ([Big97], [Obj97]). It is not meant to be complete. Missing information about the name of an approach or about its tool support means that there was no information available. It does not necessarily imply that a name or tool does not exist.Author/Institution Name Refs Type ofPub.Tool SupportAckroyd/Daum[AcDa91]A Alabisco[Ala88]CBerard[Ber93]TBailin[Bai89]ABooch[Boo94]T Y Buhr[Bu84]TCherry PAMELA 2[Che87]RCoad/Y ourdon OOA/OOD[CoYo91],[CoYo90]TTYColeman Fusion[Col94]T YAuthor/Institution Name Refs Type ofPub.Tool SupportCunningham/Beck[CuBe86]CDesfray Object Engi-neering[Des94]TEdwards Ptech[Edw89]T Y Embley et al.[EmKu92]TESA HOOD[ESA89]MFelsinger[Fel87]RFerstl/Sinz SOM[FeSi91],[FeSi90]AAYFiresmith[Fir92]TGoldberg/Rubin OBA[GoRu92]A Y Graham[Gra91]THalladay/Wiebel[HaWi93]THenderson-Sellers/Edwards[HeEd94],[Hen92]T TIBM[IBM90]RJacobson et al.OOSE[JaCh92]T Y Johnson/Foote[JoFo85]AKadie[Kad86]RKappel/Schrefl[KaSc91]ALee/Carver[LeCa91]ALiskov/Guttag[LiGu86]TMartin/Odell[MaOd92],[MaOd95]T TMasiero/Germano[MaGe88]A McGregor/Sykes[McSy92]T Meyer[Mey88]T Mullin[Mul89]T Nielsen[Nie88]R Odell[Ode92]A Page et al.[PaBe89]C Rajlich/Silva[RaSi87]R Robinson[Rob92]T2.2TerminologyOne of the shady sides of object-oriented software technology is the lack of a common, and consistent terminology. This fact may be contributed to various reasons:•The specialized research is still in a rather early stage.•There are many players - with different backgrounds and motives - who influence the evolv-ing terminology: software-engineers, computer scientists, consultants, tool vendors, etc.•There is an "inherent" ambiguity that results from the different levels of abstraction that are characteristic for the various phases of software development - resulting in subtle termino-logical traps that everybody in the field knows: Sometimes somebody, who is talking about an "object", actually means a "class", within another context he may think of a particular instance ...The definition of a sound terminology is certainly one of the key contributions to be expected from a modelling approach. For this - and for some other reasons - we do not have to discuss the whole range of terminological problems at this point. However, we do need a number of terms to categorize the approaches to be evaluated: method, methodology, model, metamodel,modelling language, etc.. Since they are essential terms not only of computer science but of science and philosophy in general, we do not intend to provide a comprehensive definition. In-stead, we will briefly describe the interpretations we favour within the context of this working paper.A method in general is a systematic approach that helps with solving a class of problems. TheAuthor/Institution Name Refs Type of Pub.Tool SupportRumbaugh et al.OMT[Rum91]T YSeidewitz/Stark [SeSt87]A Shlaer/Mellor [ShMe92], [ShMe88]T T YVelho/Carapuca SOM [VeCa92]C Walden/Nerson BON[WaNe95]T YWasserman et al.[WaPi90]A Wirfs-Brock/Wilkerson[WiWi90]TT TextbookA Article in a Journal or Reader M ManualR Research Report C Conference PaperFig. 1:Overview of past approaches to object-oriented modellingterm "systematic" is to express that usually a class of problems is divided into classes of sub-problems, together with techniques, or methods to solve them ([Lor84]). Furthermore a meth-od includes a more or less rigid temporal order of the various problem solving activities togeth-er with the required skills and resources. In general, it is a key characteristic of scientific work to apply, to review and to revise methods. A software development method is a method that is supposed to support the systematic development of a class of software.The term methodology in contrast denotes a study or a theory of methods in the sense that it includes a set of methods together with guidelines or principles to evaluate, select, apply, and develop them ([Mit84]). Many authors use "methodology" to denote a method. The methodol-ogy that is focusing on scientific methods in general, or within a particular discipline, is also called epistemology, or theory of science.A model in general is a description of an existing or potential domain, where domain denotes either a part of the "real", perceivable world, an intellectual construction, or a mixture of both. It is essential for this description that it is an abstraction of the related domain. Certain aspects of the domain are not taken into account on purpose. For this reason a model allows to concen-trate on relevant aspects, thereby helping to simplify the analysis of a complex matter. Rele-vance depends on the particular purpose a model is to serve - such as analysis, description, ex-planation, understanding, design. In the context of software development we call a model con-ceptual if it is focusing on relevant concepts of the application domain, neglecting implemen-tation-related issues. A concept is a notion, an idea that allows to identify and describe classes or categories of real world entities by defining their essential features. According to Mylopou-los and Levesque ([MyLe84], p. 11), a conceptual model should provide "... descriptions of a world enterprise/slice of reality which correspond directly and naturally to our own conceptu-alizations of the object of these descriptions.” Notice that this does not exclude models of tech-nical systems, such as machines, processors, etc. Hence, a concept is an abstraction of a set of particular occurrences. In other words: Whether or not a model is regarded as being conceptual will also depend on the intention of its creators - and the perception of the observers. Usually, but not necessarily, a conceptual model includes a graphical representation.A description requires the existence of a language. While in principle any language, in partic-ular a natural language, can be used to describe a model, there are languages that were espe-cially introduced for the purpose of describing models. We call these languages modelling lan-guages. A language may be formalized or not. There may be various representations - such as text or graphics - to express propositions of a language. Modelling languages that are used for software development usually come with a graphical notation. Languages of that kind should be suited to fulfil the requirements of conceptual modelling (focus on "natural" concepts). At the same time they should allow for a straightforward transformation to a programming lan-guage. Object-Oriented modelling is an attempt to provide modelling languages which fulfil these divergent requirements.A metalanguage is a language that serves to speak about a language, like its symbols, syntax and semantics. One inherent source of ambiguity in natural languages is the lack of a rigid sep-aration of meta- and object language. In order to avoid ambiguity, formal languages require a clear separation of meta- and object level. That does not mean, however, that a formal language and its corresponding metalanguage need to be different in terms of their symbols or their syn-tax. Defining a language with a formal metalanguage promises a number of benefits. A meta-language that is used to specify a number of object-level languages helps to integrate these lan-guages: Information represented in the various object-level languages can be exchanged by re-ferring to concepts defined in the common metalanguage - as long as the involved languages have those concepts in common. Furthermore a metalanguage can help with reusability and flexibility: Software that operates on a particular metalanguage, can be used to generate soft-ware that operates on the object-level language (like a compiler compiler, or Meta CASE, see for instance [EbSü96]). A metalanguage may be described by yet another metalanguage which is sometimes called meta metalanguage, and so forth. A meta metalanguage provides the same advantages for handling metalanguages as a metalanguage does for object-level languages.A metamodel is a model representing a metalanguage of a class of modelling languages. It serves to describe the symbols, syntax, and semantics of these modelling languages in a more or less rigid way. Depending on the metalanguage it is based on a metamodel may be formal or not. A metamodel may also include instructions or recommendations of how to render a model for the viewer, in other words: the notation of the modelling language. A metamodel may be rendered in the same notation as its subordinated models. The benefits provided by metamodels or meta metamodels correspond to those of metalanguages and meta metalan-guages respectively.A modelling method is a method that supports the construction and use of a class of models. It will usually include some sort of metamodel or metalanguage that, however, does not have to be formalized. Furthermore it may include some of the following aspects (this list is not meant to be complete):•more or less detailed decomposition into subproblems together with problem solving activ-ities•temporal/causal order of problem solving activities•profiles of roles required to perform those activities•information required, together with heuristics to gather it•documents/models to be produced•criteria or even metrics to evaluate models•further support for project management•examples of how to apply the method•generic models for certain domains that can be reused and modifiedA modelling methodology is a study or a theory about modelling methods within a certain area. Its subject is a set of modelling methods together with criteria/principles to use, analyze, and revise them. Computer science in general could be regarded as a modelling methodology, since one of its essential research goals is to provide appropriate abstractions for a class of problems - together with solutions to these problems.Fig. 2 shows a semantic net that illustrates the terminology. It is not intended to provide a com-plete and precise definition.2.3Subjects, Benefits, and Pitfalls of StandardizationWhile conventions and standards have played an important role in information technology for long, standards have gained a still increasing attention in the nineties. It is remarkable that to-day’s information managers are not only interested in suitable standards. Instead they are con-Fig. 2:Concepts within object-oriented modellingMethodMethodologyNotation Language Model Metalanguage MetamodelModellingLanguageProblem Heuristics ModellingMethodology ModellingMethodDomain Natural Language Level of Formalization Technique Processis a Artificial Language is acharacterized byrequiresserves to express serves to describe is adescribes/definesserves asserves to expressrepresents includes +++++includes+includesis a includesincludes includes should help with solving +is subject of ++is subject ofis a+++more than one possible serves ascentrating more than ever on the question which attempt will finally make it as the one and only global standard for a certain area - thereby almost paralyzing their ability to make deci-sions. We believe that this is both for economic and psychological reasons. The emergence of microcomputers and so called industry standards (which are sometimes in fact proprietary specifications) have demonstrated the tremendous impact of scale economies on the produc-tion costs per unit of hardware, and especially of software. This experience seems to have fos-tered a wide spread believe that in the end only one standard within a particular scope will sur-vive. Not only vendors, but also (organizational) buyers are obviously impressed (not to say "obsessed") with this "the winner takes it all" mentality. From an economic point of view it makes a big difference, whether a "standard" results from a dominating proprietary technolo-gy, or whether it is "open" in the sense that its specification may - in principle - be accessed, and used by everybody. The specification of open standards has been the business of national or international standardization bodies for long. There have also been standardization initia-tives which were organized by the industry itself, mainly for two reasons: to speed up the standardization process and to gain better control over it. Before we will have a look at one particular initiative that is of outstanding importance within the area of object-oriented tech-nologies, namely the OMG, we will first analyze what could be subject of standardization in the area of object-oriented modelling.Standardizing a (Modelling) MethodologyA methodology in the sense of how we introduced the term is by no means a subject suitable for standardization. It is common sense that a theory should be permanently evaluated against the domain it has been introduced for - and revised if necessary. This epistemological process would be compromised by "standard theories". Standardizing a study would be equally bizarre. There is, however, no doubt that a common terminology is rather important for the evolution of methodologies as well.Standardizing a (Modelling) MethodSimilar to methodology it is not obvious, why it would make sense to standardize "a systematic approach that helps with solving a class of problems". Instead it seems to be more appropriate that anybody, who has to solve a problem, may try the available methods and finally take the one that suits him best. From an academic point of view such a standardization seems to be bizarre: Scientific progress relies heavily on permanent reviews and - eventually - modifica-tions of existing methods. Freezing a method by standardizing it would contradict this process. However, while this is certainly true from an epistemological point of view, things tend to be more subtle in practice. For one, we have to take into account that defining and prescribing sta-ble methods - with more or less rigour - is an essential prerequisite of professional collabora-tive work ([Wei79], pp. 15). Organizing allows for task specific training, and is a prerequisite for core management functions such as planning and control. Does that mean, however, that there is need to standardize business processes? There is certainly a clear objection against standardizing the organisation of processes: Organizing its business should be part of the core competencies of almost any company, and therefore it is an essential part of its competitive-ness. On the other hand we currently witness a tendency toward standard business processes, promoted by technologies such as integrated business software, or workflow management sys-tems. Reorganising business processes by applying certain standard patterns facilitates the ef-ficient introduction of corresponding software. The question of the appropriate level of stand-ardization within business processes is a complex matter that we cannot analyze here in depth.There is, however, another motivation for standardizing methods: Defining steps and guide-lines for documenting their results may help with reconstructing a business process. Thereby customers may get an idea of the efficiency of a process and how the process contributes to the overall quality of a product or service. For the companies that apply the method it may help with increasing the awareness of organizational efficiency and thereby fostering more efficient processes. Recent examples of efforts to standardize methods or processes are ISO 9000 ([KeJa96], [Sch94]) and the "V-Modell" (a process model for software development, defined by the German forces "Bundeswehr" [BrDr93]). Notice that standard methods such as ISO 9000 or the "V-Modell" define a temporal order of documents and their corresponding (mod-elling) languages (which may be structured forms of the natural language) rather than provid-ing a precise definition of the involved tasks and how to perform them.We can summarize that standardizing a software development method in every detail is cer-tainly no option. A method’s potential depends on the problem it is used for, the skills and at-titude of the method users, the available resources, etc. For this reason it is hard to imagine that there could be one best (object-oriented modelling) method at a point in time. Furthermore you have to take into account that methods will (and should) evolve over time. At last it would cer-tainly be too expensive to enforce a detailed standard method, since that would cause expenses for monitoring which nobody could afford. Therefore it is definitely no good idea to standard-ize a method. However, in order to foster comparability, quality, and professional training, it can make sense to standardize the documents (models) that should be produced within a meth-od - which in turn will certainly have an impact on the methods that use these documents. Standardizing a (Modelling) LanguageThe notion of a language implies the idea of conventions which may be more or less rigid. Oth-erwise a language could not be used for the purpose of communication. Standardizing a lan-guage implies its formalization which in turn requires a formal metalanguage. It is evident that the metalanguage itself has to be precisely defined, too. In order to avoid a regressum ad in-finitum there has to be at one point a reference to already existing well defined languages and concepts - like a logical calculus or standardized types/classes.What are the benefits of standardizing modelling languages used within the process of soft-ware development? The potential benefits are obvious: Standardized modelling languages will allow for better protection of investments in technologies that depend on those languages. This is also the case for investments into training. Standardized modelling language will also im-prove the chance for exchanging models and interoperability in general, or - in other words -for system integration. Furthermore standardized modelling languages will foster reusability of existing models, for instance of generic or reference models developed for certain problem areas or domains. A modelling language will usually include a graphical notation which, how-ever, does not have to be part of a standard: From a logical point of view the notation is irrel-evant, since it does not contain any formal semantics. Nevertheless a common notation will facilitate the communication between all people using the modelling language. Therefore it might be reasonable to enhance a standard with recommendations for a notation.Despite its potential benefits, standardizing modelling languages faces a severe challenge: What are the requirements a modelling language should fulfil in order to be useful within a longer time frame? We will analyze this question later (see 4.3, [FrPr97]). One thing, however, is for sure: There is no easy answer to it. This leads to another question we will also look at later in this report: Is it already the time to standardize object-oriented modelling languages?。
- 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)。
《纳氏英文文法》第二册阅读趣话
英国学者—— 纳斯斐尔德（J. C. Nesfield, M. A.），生卒年不详。 英国语言学家，曾获文科硕士学位，在印度从事过教育活动，主要 以研究实用英语和作文文法等见长，他的大多数著作都是为非英语 国家的英语学习者撰写的，曾著有：
英语原著欣赏
——《纳氏英文文法》第二册阅读趣话
《纳氏英文文法》第二册阅读趣话
一、淘 书
《纳氏英文文法》第二册阅读趣话
English Grammar Series Book II Easy Parsing & Analysis
With Chinese Translation
J. C. Nesfield, M.A.
《成语、文法与综合》（Idiom,Grammar, and Synthesis, 1895） 《英语语法与作文手册》（Manual of English Grammar and Composition,
1898）
《英语语法的过去和现在》（English Grammar, Past and Present, 1898） 《英语作文的错误》（Errors in English Composition, 1903） 《历史上的英语和起源》（Historical English and Derivation）等等。
《纳氏英文文法》第二册阅读趣话
澳大利亚国家图书馆回函一（08-12-1）： We expect to respond by: 08/12/08 02:40 pm We will send our response to the following email: yu_wr2019yahoo. We noticed that the question's last useful date to you is: 08/12/08 02:45 pm We will make every effort to respond to your question within one week. However, complex enquiries relating to manuscripts, music, oral history or pictures may take up to four weeks. Thank you for using our service. National Library of Australia
《纳氏英文文法》第二册阅读趣话
澳大利亚国家图书馆回函二（08-12-3）：
William Thank you for your query. Unfortunately although I found many references to books by John Collinson Nesfield I have found no references to his actual life, other than that he had lived for many years in India (he was an inspector of schools in Oudh) and his first grammar books were written for the Indian market. I have found some birth and death dates but these are uncertain, as the reference I have says only that he 'flourished' (fl) between 1885-1949.
《纳氏英文文法》第二册阅读趣话 （封 面）
（书 名）
《纳氏英文文法》第二册阅读趣话
（目 录） 底）
（封
《纳氏英文文法》第二册阅读趣话
二、研 读
《纳氏英文文法》第二册阅读趣话
相关信息：
※ 《纳氏英文文法》：简称《纳氏文法》，全套四册共有几十本之多。 ※ J. C. Nesfield：全名John Collinson Nesfield（纳斯菲尔德），英
国著名语法专家。 ※ 麦克米兰出版公司：Macmillan Publishers Ltd. ※ 出版时间：1895年第一版，之后曾再版不少于30次。 ※ 该书译者：伦敦大学 徐兆熊（英汉本出版于1921年），留英学者。 ※ １９２６年，浮石人赵灼出版《纳氏英文法讲义》，这是中国第一本
英文语法书，用中文和英文讲解英文语法，全书四大册，是当时全中 国各大学都采用的教材。——江门日报（广东省）
《纳氏英文文法》第二册阅读趣话
英国麦克米兰出版公司回函（08-1-11）：
From: macmillaneducationmacmillan To: waijiaozhongxinhotmail Date: Fri, 11 Jan 2019 08:20:29 +0000 Subject: Macmillan Education Contact Form Thank you for contacting macmillaneducation Your enquiry has been forwarded to the relevant department and will be dealt with as soon as possible. Best wishes Macmillan Education Website Editorial Team macmillaneducation
《纳氏英文文法》第二册阅读趣话
广告词语：
納氏
英文法講義
趙 灼 编译
卷一
全一冊 定價大洋 六 角
卷二
全一冊 定價大洋 一 元
卷三
全一冊
定價大洋 二元四角
卷四
全一冊 定價大洋 五 元
納士斐而文典 Nesfield‘s English Grammar Series，為英文新著 中最
善之作。近年以來，我國校率皆用之以作教本。惟原書係教科