Demand-driven type analysis for dynamically-typed functional languages

Thèse diffusée initialement dans le cadre d'un projet pilote des Presses de l'Université de Montréal/Centre d'édition numérique UdeM (1997-2008) avec l'autorisation de l'auteur.

Using synoptic type analysis to understand New Zealand climate during the Mid-Holocene

Diagnosing the climate of New Zealand from low-resolution General Circulation Models (GCMs) is notoriously difficult due to the interaction of the complex topography and the Southern Hemisphere (SH) mid-latitude westerly winds. Therefore, methods of downscaling synoptic scale model data for New Zealand are useful to help understand past climate. New Zealand also has a wealth of palaeoclimate-proxy data to which the downscaled model output can be compared, and to provide a qualitative method of assessing the capability of GCMs to represent, in this case, the climate 6000 yr ago in the Mid-Holocene. In this paper, a synoptic weather and climate regime classification system using Empirical Orthogonal Function (EOF) analysis of GCM and reanalysis data was used. The climate regimes are associated with surface air temperature and precipitation anomalies over New Zealand. From the analysis in this study, we find at 6000 BP that increased trough activity in summer and autumn led to increased precipitation, with an increased north-south pressure gradient ("zonal events") in winter and spring leading to drier conditions. Opposing effects of increased (decreased) temperature are also seen in spring (autumn) in the South Island, which are associated with the increased zonal (trough) events; however, the circulation induced changes in temperature are likely to have been of secondary importance to the insolation induced changes. Evidence from the palaeoclimate-proxy data suggests that the Mid-Holocene was characterized by increased westerly wind events in New Zealand, which agrees with the preference for trough and zonal regimes in the models.

Inferring determinacy and mutual exclusion in logic programs using mode and type analysis.

We propose an analysis for detecting procedures and goals that are deterministic (i.e., that produce at most one solution at most once), or predicates whose clause tests are mutually exclusive (which implies that at most one of their clauses will succeed) even if they are not deterministic. The analysis takes advantage of the pruning operator in order to improve the detection of mutual exclusion and determinacy. It also supports arithmetic equations and disequations, as well as equations and disequations on terms, for which we give a complete satisfiability testing algorithm, w.r.t. available type information. We have implemented the analysis and integrated it in the CiaoPP system, which also infers automatically the mode and type information that our analysis takes as input. Experiments performed on this implementation show that the analysis is fairly accurate and efficient.

Sized Type Analysis for Logic Programs (Technical Communication)

We present a novel analysis for relating the sizes of terms and subterms occurring at diferent argument positions in logic predicates. We extend and enrich the concept of sized type as a representation that incorporates structural (shape) information and allows expressing both lower and upper bounds on the size of a set of terms and their subterms at any position and depth. For example, expressing bounds on the length of lists of numbers, together with bounds on the values of all of their elements. The analysis is developed using abstract interpretation and the novel abstract operations are based on setting up and solving recurrence relations between sized types. It has been integrated, together with novel resource usage and cardinality analyses, in the abstract interpretation framework in the Ciao preprocessor, CiaoPP, in order to assess both the accuracy of the new size analysis and its usefulness in the resource usage estimation application. We show that the proposed sized types are a substantial improvement over the previous size analyses present in CiaoPP, and also benefit the resource analysis considerably, allowing the inference of equal or better bounds than comparable state of the art systems.

On the fly type specialization without type analysis

Les langages de programmation typés dynamiquement tels que JavaScript et Python repoussent la vérification de typage jusqu’au moment de l’exécution. Afin d’optimiser la performance de ces langages, les implémentations de machines virtuelles pour langages dynamiques doivent tenter d’éliminer les tests de typage dynamiques redondants. Cela se fait habituellement en utilisant une analyse d’inférence de types. Cependant, les analyses de ce genre sont souvent coûteuses et impliquent des compromis entre le temps de compilation et la précision des résultats obtenus. Ceci a conduit à la conception d’architectures de VM de plus en plus complexes. Nous proposons le versionnement paresseux de blocs de base, une technique de compilation à la volée simple qui élimine efficacement les tests de typage dynamiques redondants sur les chemins d’exécution critiques. Cette nouvelle approche génère paresseusement des versions spécialisées des blocs de base tout en propageant de l’information de typage contextualisée. Notre technique ne nécessite pas l’utilisation d’analyses de programme coûteuses, n’est pas contrainte par les limitations de précision des analyses d’inférence de types traditionnelles et évite la complexité des techniques d’optimisation spéculatives. Trois extensions sont apportées au versionnement de blocs de base afin de lui donner des capacités d’optimisation interprocédurale. Une première extension lui donne la possibilité de joindre des informations de typage aux propriétés des objets et aux variables globales. Puis, la spécialisation de points d’entrée lui permet de passer de l’information de typage des fonctions appellantes aux fonctions appellées. Finalement, la spécialisation des continuations d’appels permet de transmettre le type des valeurs de retour des fonctions appellées aux appellants sans coût dynamique. Nous démontrons empiriquement que ces extensions permettent au versionnement de blocs de base d’éliminer plus de tests de typage dynamiques que toute analyse d’inférence de typage statique.

On the fly type specialization without type analysis

Les langages de programmation typés dynamiquement tels que JavaScript et Python repoussent la vérification de typage jusqu’au moment de l’exécution. Afin d’optimiser la performance de ces langages, les implémentations de machines virtuelles pour langages dynamiques doivent tenter d’éliminer les tests de typage dynamiques redondants. Cela se fait habituellement en utilisant une analyse d’inférence de types. Cependant, les analyses de ce genre sont souvent coûteuses et impliquent des compromis entre le temps de compilation et la précision des résultats obtenus. Ceci a conduit à la conception d’architectures de VM de plus en plus complexes. Nous proposons le versionnement paresseux de blocs de base, une technique de compilation à la volée simple qui élimine efficacement les tests de typage dynamiques redondants sur les chemins d’exécution critiques. Cette nouvelle approche génère paresseusement des versions spécialisées des blocs de base tout en propageant de l’information de typage contextualisée. Notre technique ne nécessite pas l’utilisation d’analyses de programme coûteuses, n’est pas contrainte par les limitations de précision des analyses d’inférence de types traditionnelles et évite la complexité des techniques d’optimisation spéculatives. Trois extensions sont apportées au versionnement de blocs de base afin de lui donner des capacités d’optimisation interprocédurale. Une première extension lui donne la possibilité de joindre des informations de typage aux propriétés des objets et aux variables globales. Puis, la spécialisation de points d’entrée lui permet de passer de l’information de typage des fonctions appellantes aux fonctions appellées. Finalement, la spécialisation des continuations d’appels permet de transmettre le type des valeurs de retour des fonctions appellées aux appellants sans coût dynamique. Nous démontrons empiriquement que ces extensions permettent au versionnement de blocs de base d’éliminer plus de tests de typage dynamiques que toute analyse d’inférence de typage statique.

Genetic structure and mating type analysis of the Pyricularia oryzae population causing widespread epidemics in southern Brazil.

Blast is a major disease of rice in Brazil, the largest rice-producing country outside Asia. This study aimed to assess the genetic structure and mating-type frequency in a contemporary Pyricularia oryzae population, which caused widespread epidemics during the 2012/13 season in the Brazilian lowland subtropical region. Symptomatic leaves and panicles were sampled at flooded rice fields in the states of Rio Grande do Sul (RS, 34 fields) and Santa Catarina (SC, 21 fields). The polymorphism at ten simple sequence repeats (SSR or microsatellite) loci and the presence of MAT1-1 or MAT1-2 idiomorphs were assessed in a population comprised of 187 isolates. Only the MAT1-2 idiomorph was found and 162 genotypes were identified by the SSR analysis. A discriminant analysis of principal components (DAPC) of SSR data resolved four genetic groups, which were strongly associated with the cultivar of origin of the isolates. There was high level of genotypic diversity and moderate level of gene diversity regardless whether isolates were grouped in subpopulations based on geographic region, cultivar host or cultivar within region. While regional subpopulations were weakly differentiated, high genetic differentiation was found among subpopulations comprised of isolates from different cultivars. The data suggest that the rice blast pathogen population in southern Brazil is comprised of clonal lineages that are adapting to specific cultivar hosts. Farmers should avoid the use of susceptible cultivars over large areas and breeders should focus at enlarging the genetic basis of new cultivars.

Moving Beyond Teen Crash Fatality Statistics: The Go-Team Study, GOTM-000, 2013

Despite a trend of decreasing teen fatalities due to motor vehicle crashes over the past decade, they remain the leading cause of adolescent fatalities in Iowa. The purpose of this study was to create detailed case studies of each fatal motor vehicle crash involving a driver under the age of 20 that occurred in Iowa in 2009, 2010, and 2011. Data for each crash were gathered from media sources, law enforcement agencies, and the Iowa Department of Transportation. The driving records of the teens, which included their licensure history, prior traffic citations, and prior crashes, were also acquired. In addition, data about the charges filed against a teen as a result of being involved in a fatal crash were obtained. A total of 126 crashes involving 131 teen drivers that resulted in 143 fatalities were analyzed. Many findings for fatal crashes involving teen drivers in Iowa are consistent with national trends, including the overrepresentation of male drivers, crash involvement that increases with age, crash involvement per vehicle miles traveled that decreases with age, and prevalence of single-vehicle road departure crashes. Relative to national statistics, teen fatalities from crashes in Iowa are more likely to occur from midnight to 6am and from 9am to noon. Crash type varied by driver age and county population level. Teen drivers contributed to the fatal crashes at a rate of 74%; contribution of the teen driver was unknown for 11% of crashes. Speed was a factor for about 25% of the crashes for which a teen driver was at fault. The same was also true of alcohol/drug impairment. Only 20% of the rear-seat occupants of the teen drivers’ vehicles wore seat belts compared to 60% use for the front-seat occupants. Analysis of the teens’ driving records prior to the fatal crash suggests at-fault crashes and speeding violations are associated with contributing to the fatal crash.

Spatial Scale of Clustering of Motor Vehicle Crash Types and Appropriate Countermeasures, 2009

This project analyzes the characteristics and spatial distributions of motor vehicle crash types in order to evaluate the degree and scale of their spatial clustering. Crashes occur as the result of a variety of vehicle, roadway, and human factors and thus vary in their clustering behavior. Clustering can occur at a variety of scales, from the intersection level, to the corridor level, to the area level. Conversely, other crash types are less linked to geographic factors and are more spatially “random.” The degree and scale of clustering have implications for the use of strategies to promote transportation safety. In this project, Iowa's crash database, geographic information systems, and recent advances in spatial statistics methodologies and software tools were used to analyze the degree and spatial scale of clustering for several crash types within the counties of the Iowa Northland Regional Council of Governments. A statistical measure called the K function was used to analyze the clustering behavior of crashes. Several methodological issues, related to the application of this spatial statistical technique in the context of motor vehicle crashes on a road network, were identified and addressed. These methods facilitated the identification of crash clusters at appropriate scales of analysis for each crash type. This clustering information is useful for improving transportation safety through focused countermeasures directly linked to crash causes and the spatial extent of identified problem locations, as well as through the identification of less location-based crash types better suited to non-spatial countermeasures. The results of the K function analysis point to the usefulness of the procedure in identifying the degree and scale at which crashes cluster, or do not cluster, relative to each other. Moreover, for many individual crash types, different patterns and processes and potentially different countermeasures appeared at different scales of analysis. This finding highlights the importance of scale considerations in problem identification and countermeasure formulation.

Floquet stability analysis of the flow around an oscillating cylinder

This investigative work is concerned with the flow around a circular cylinder submitted to forced transverse oscillations. The goal is to investigate how the transition to turbulence is initiated in the wake for cases with different Reynolds numbers (Re) and displacement amplitudes (A). For each Re the motion frequency is kept constant, close to the Strouhal number of the flow around a fixed cylinder at the same Re. Stability analysis of two-dimensional periodic flows around a forced-oscillating cylinder is carried out with respect to three-dimensional infinitesimal perturbations. The procedure consists of performing a Floquet type analysis of time-periodic base flows, computed using the spectral/hp element method. With the results of the Floquet calculations, considerations regarding the stability of the system are drawn, and the form of the instability at its onset is obtained. The critical Reynolds number is observed to change with the amplitude of oscillation. With respect to instabilities, unstable modes with the same symmetry as mode A of a fixed cylinder are observed; however, they present different wavelengths. Also, the instabilities observed for the oscillating cylinder are distinctively stronger in the braid shear layers. Other unstable modes similar to mode B are found. Quasi-periodic modes are observed in the 2S wake, and subharmonic mode occurrences are reported in P + S wakes. (C) 2009 Elsevier Ltd. All rights reserved.

Genotyping, serotyping and determination of mating-type of Cryptococcus neoformans clinical isolates from São Paulo State, Brazil

The basidiomycetous yeast Cryptococcus neoformans is an important fungal pathogen mainly in immunocompromised patients. In this study, 47 clinical isolates of C. neoformans from regions of São Paulo State were studied serologically by using the Crypto Check Iatron RM 304-K kit, their genetic diversity was estimated by PCR-fingerprinting with a microsatellite-specific sequence (GACA)4, RAPD with primer 6 (Amersham Pharmacia Biotech), PCR-restriction fragment length polymorphism (RFLP) analysis of the phospholipase B gene (PLB1) digested with AvaI and mating type analysis by PCR. All 47 strains isolated from HIV positive patients included in this study were serotype A and MATalpha. The majority of the isolates (45/47) were VNI and only two were VNII by PCR-fingerprinting and PCR-RFLP analysis. High degree of homogeneity was observed when (GACA)4 was used, being highly correlated (> 0.9). In contrast, the RAPD analysis was more heterogeneous with higher number of molecular profiles. By PCR-RFLP, no new molecular type was found, enhancing the suggestion that the differences based on conserved gene as PLB1, can be resultant of ongoing divergent evolution within the C. neoformans complex, into the current eight subtypes. Our results furnish new information on the molecular epidemiology of C. neoformans in the southeast region of Brazil.

Guidelines for the Conversion of Urban Four-Lane Undivided Roadways to Three-Lane Two-Way Left-turn Lane Facilities, April 2001

Four-lane undivided roadways in urban areas can experience a degradation of service and/or safety as traffic volumes increase. In fact, the existence of turning vehicles on this type of roadway has a dramatic effect on both of these factors. The solution identified for these problems is typically the addition of a raised median or two-way left-turn lane (TWLTL). The mobility and safety benefits of these actions have been proven and are discussed in the “Past Research” chapter of this report along with some general cross section selection guidelines. The cost and right-of-way impacts of these actions are widely accepted. These guidelines focus on the evaluation and analysis of an alternative to the typical four-lane undivided cross section improvement approach described above. It has been found that the conversion of a four-lane undivided cross section to three lanes (i.e., one lane in each direction and a TWLTL) can improve safety and maintain an acceptable level of service. These guidelines summarize the results of past research in this area (which is almost nonexistent) and qualitative/quantitative before-and-after safety and operational impacts of case study conversions located throughout the United States and Iowa. Past research confirms that this type of conversion is acceptable or feasible in some situations but for the most part fails to specifically identify those situations. In general, the reviewed case study conversions resulted in a reduction of average or 85th percentile speeds (typically less than five miles per hour) and a relatively dramatic reduction in excessive speeding (a 60 to 70 percent reduction in the number of vehicles traveling five miles per hour faster than the posted speed limit was measured in two cases) and total crashes (reductions between 17 to 62 percent were measured). The 13 roadway conversions considered had average daily traffic volumes of 8,400 to 14,000 vehicles per day (vpd) in Iowa and 9,200 to 24,000 vehicles per day elsewhere. In addition to past research and case study results, a simulation sensitivity analysis was completed to investigate and/or confirm the operational impacts of a four-lane undivided to three-lane conversion. First, the advantages and disadvantages of different corridor simulation packages were identified for this type of analysis. Then, the CORridor SIMulation (CORSIM) software was used x to investigate and evaluate several characteristics related to the operational feasibility of a four-lane undivided to three-lane conversion. Simulated speed and level of service results for both cross sections were documented for different total peak-hour traffic, access densities, and access-point left-turn volumes (for a case study corridor defined by the researchers). These analyses assisted with the identification of the considerations for the operational feasibility determination of a four -lane to three-lane conversion. The results of the simulation analyses primarily confirmed the case study impacts. The CORSIM results indicated only a slight decrease in average arterial speed for through vehicles can be expected for a large range of peak-hour volumes, access densities, and access-point left-turn volumes (given the assumptions and design of the corridor case study evaluated). Typically, the reduction in the simulated average arterial speed (which includes both segment and signal delay) was between zero and four miles per hour when a roadway was converted from a four-lane undivided to a three-lane cross section. The simulated arterial level of service for a converted roadway, however, showed a decrease when the bi-directional peak-hour volume was about 1,750 vehicles per hour (or 17,500 vehicles per day if 10 percent of the daily volume is assumed to occur in the peak hour). Past research by others, however, indicates that 12,000 vehicles per day may be the operational capacity (i.e., level of service E) of a three-lane roadway due to vehicle platooning. The simulation results, along with past research and case study results, appear to support following volume-related feasibility suggestions for four-lane undivided to three-lane cross section conversions. It is recommended that a four-lane undivided to three-lane conversion be considered as a feasible (with respect to volume only) option when bi-directional peak-hour volumes are less than 1,500 vehicles per hour, but that some caution begin to be exercised when the roadway has a bi-directional peak-hour volume between 1,500 and 1,750 vehicles per hour. At and above 1,750 vehicles per hour, the simulation indicated a reduction in arterial level of service. Therefore, at least in Iowa, the feasibility of a four-lane undivided to three-lane conversion should be questioned and/or considered much more closely when a roadway has (or is expected to have) a peak-hour volume of more than 1,750 vehicles. Assuming that 10 percent of the daily traffic occurs during the peak-hour, these volume recommendations would correspond to 15,000 and 17,500 vehicles per day, respectively. These suggestions, however, are based on the results from one idealized case xi study corridor analysis. Individual operational analysis and/or simulations should be completed in detail once a four-lane undivided to three-lane cross section conversion is considered feasible (based on the general suggestions above) for a particular corridor. All of the simulations completed as part of this project also incorporated the optimization of signal timing to minimize vehicle delay along the corridor. A number of determination feasibility factors were identified from a review of the past research, before-and-after case study results, and the simulation sensitivity analysis. The existing and expected (i.e., design period) statuses of these factors are described and should be considered. The characteristics of these factors should be compared to each other, the impacts of other potentially feasible cross section improvements, and the goals/objectives of the community. The factors discussed in these guidelines include • roadway function and environment • overall traffic volume and level of service • turning volumes and patterns • frequent-stop and slow-moving vehicles • weaving, speed, and queues • crash type and patterns • pedestrian and bike activity • right-of-way availability, cost, and acquisition impacts • general characteristics, including - parallel roadways - offset minor street intersections - parallel parking - corner radii - at-grade railroad crossings xii The characteristics of these factors are documented in these guidelines, and their relationship to four-lane undivided to three-lane cross section conversion feasibility identified. This information is summarized along with some evaluative questions in this executive summary and Appendix C. In summary, the results of past research, numerous case studies, and the simulation analyses done as part of this project support the conclusion that in certain circumstances a four-lane undivided to three-lane conversion can be a feasible alternative for the mitigation of operational and/or safety concerns. This feasibility, however, must be determined by an evaluation of the factors identified in these guidelines (along with any others that may be relevant for a individual corridor). The expected benefits, costs, and overall impacts of a four-lane undivided to three-lane conversion should then be compared to the impacts of other feasible alternatives (e.g., adding a raised median) at a particular location.

Evaluation of Rumble Stripes on Low-Volume Rural Roads in Iowa—Phase II Final Report, November 2011

Single-vehicle run-off-road crashes are the most common crash type on rural two-lane Iowa roads. Rumble strips have proven effective in mitigating these crashes, but the strips are commonly installed in paved shoulders on higher-volume roads that are owned by the State of Iowa. Lower-volume paved rural roads owned by local agencies do not commonly feature paved shoulders but frequently experience run-off-road crashes. This project involved installing rumble stripes, which are a combination of conventional rumble strips with a painted edge line placed on the surface of the milled area, along the edge of the travel lanes, but at a narrow width to avoid possible intrusion into the normal vehicle travel paths. The research described in this report was part of a project funded by the Federal Highway Administration, Iowa Highway Research Board, and Iowa Department of Transportation to evaluate the effectiveness of edge-line rumble strips in Iowa. The project evaluated the effectiveness of rumble stripes in reducing run-off-road crashes and in improving the longevity and wet-weather visibility of edge-line markings. This project consisted of two phases. The first phase was to select pilot study locations, select a set of test sites, install rumble stripes, summarize lessons learned during installation, and provide a preliminary assessment of the rumble stripes’ performance. The purpose of this report was to document results from Phase II. A before and after crash analysis was conducted to assess whether use of the treatment had resulted in fewer crashes. However, due to low sample size, results of the analysis were inconclusive. Lateral position was also evaluated before and after installation of the treatment to determine whether vehicles engaged in better lane keeping. Pavement marking wear was also assessed.

Genotyping, serotyping and determination of mating-type of Cryptococcus neoformans clinical isolates from São Paulo State, Brazil

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)

Neuropathological heterogeneity in Alzheimer's disease:A study of 80 cases using principal components analysis

Three hypotheses have been proposed to explain neuropathological heterogeneity in Alzheimer's disease (AD): the presence of distinct subtypes ('subtype hypothesis'), variation in the stage of the disease ('phase hypothesis') and variation in the origin and progression of the disease ('compensation hypothesis'). To test these hypotheses, variation in the distribution and severity of senile plaques (SP) and neurofibrillary tangles (NFT) was studied in 80 cases of AD using principal components analysis (PCA). Principal components analysis using the cases as variables (Q-type analysis) suggested that individual differences between patients were continuously distributed rather than the cases being clustered into distinct subtypes. In addition, PCA using the abundances of SP and NFT as variables (R-type analysis) suggested that variations in the presence and abundance of lesions in the frontal and occipital lobes, the cingulate gyrus and the posterior parahippocampal gyrus were the most important sources of heterogeneity consistent with the presence of different stages of the disease. In addition, in a subgroup of patients, individual differences were related to apolipoprotein E (ApoE) genotype, the presence and severity of SP in the frontal and occipital cortex being significantly increased in patients expressing apolipoprotein (Apo)E allele ε4. It was concluded that some of the neuropathological heterogeneity in our AD cases may be consistent with the 'phase hypothesis'. A major factor determining this variation in late-onset cases was ApoE genotype with accelerated rates of spread of the pathology in patients expressing allele ε4.