Author: Lars Holm Nielsen
Language: English
Year: 2007
The abstract and PDF for this report is not public available as the project was made at IBM Research in Zürich.
Keywords: Bayesian networks, probability elicitation, expert knowledge, probability aggregation, questionnaires, artifical intelligence, reasoning.
Authors: Steffen Dreyer Andersen & Lars Holm Nielsen
Language: Danish
Year: 2006
This report treats the implementation and integration of program analysis in Eclipse. We have in the project on which this report is based, developed two plug-ins, that makes it possible to construct interprocedural programanalysis. The development had its origen in the implementation of a programanalysis for detecting unreachable methodes i Java. The development process and working with Eclipse has given us an insite to the possibilities for integration and implementing programanalysis in Eclipse. The conclusion is that there are good possibilities for integrating programanalysis i Eclipse. The TPTP platform makes it possible to focus on the implementation of the analysis instead of Eclipse integration. We have with our plug-ins showed that it is possible to implement an interprocedural analysis although TPTP is targeted at that.
Keywords: Program analysis, Eclipse, Integrated Development Environments, Java.
Author: Lars Holm Nielsen
Language: Danish
Year: 2006
Set-based program analysis of logic programs basically consists of constraint generation and constraint resolution. The algorithms for solving systems of set constraints are based on exhaustively application of transformations to the system of set constraints. The solution of a system of set constraints are however NEXPTIME-complete, and thus it can largely be intractable to find a solution. In this report, we present a naive implementation of a know algorithm to solving systems of set constraints. Based on the implementation, we have made experiments with both a Prolog program and constructed set constraints. The experiments have shown, that the primary cause of intractability is partly an explosive increase in constraints, caused by three parameters, and partly the number of iterations where especially the last iteration is very costly but however highly unproductive. Furthermore there have been shown a large redundancy in the solution to systems of set constraints.
Keywords: Program analysis, set-based analysis, set constraints, algorithms, predicate logic.
Authors: Lars Holm Nielsen, Nanna Torpe Jørgensen, Kim Jantzen and Sanne Bjerg
Language: English
Year: 2005
Twelve qualitative open-ended in-depth interviews with public information officers from large governmental scientific institutions as well as scientists and journalist were conducted in November 2005. An analysis of the interviews is carried out, to study how far science communicators in the name of science communication can, or should, push, without damaging the individual, and thus also the collective credibility, of the science communication community and the involved institutions. Overall the study suggests that a credibility problem for astronomical press releases does not exist. However the interviews indicated a problem within the scientific community, where some lack of understanding and respect among the actors is found. All actors though, showed great attention and concern for communicating science in a credible way. A code of conduct presenting guidelines for how to minimize hype in press releases is presented.
Keywords: credibility, hype, science communication, astronomy, visibility and press releases.
Authors: Lars Holm Nielsen, Regin Gerner Hansen, Andreas Erland Jensen and Tonny Elgaard Larsen
Language: Danish
Year: 2005
Compilers for programming languages such as C/C++ and Java have a built-in static type check. It’s common to use this kind of type check to analyze programs for type errors. The purpose of the static type check is that typed programming languages contribute to less the occurrence of less run- time errors, compared to untyped programming languages. In this report, we study to what extent static type check can be used for untyped script languages. The study has been conducted by making a program, which make a static type check of PHP-programs. The program has been tested with a number of test programs, to expose strengths and weaknesses in static type check of untyped script languages. We conclude that a static type check can be used to analyze script languages with dynamic type check for type errors, to the same extent as static type check can be used to for typed languages. However, this happens at the expense of flexibility in the untyped script languages.
Keywords: computer science, types, type check, type inference, type reconstruction, soft types, type systems, compilers, programming languages, PHP
Authors: Lars Holm Nielsen, Kim Jantzen, Anders Rohde, June Liljefeld Nielsen, Christian Schmidt and Jens Bang Mikkelsen
Language: Danish
Year: 2004
Trehalose is a nonreducing disaccharide of glucose. Scientist from Cornell University has created a stress tolerant transgenic rice with increased accumulation of trehalose. We report here how scientist from the Cornell University have engineered a rice trait, and we examine if the utilization of the rice trait is favorable regarding increased production of food. We describe the lack of risks associated with digestion of trehalose. However thorough assessment of the consequences of gene flow by hybridization from transgenic rice is needed. We conclude, that the utilization of the stress tolerant transgenic rice will be favorable as long as sufficient precautions regarding the gene flow are made.
Keywords: Transgenic rice, trehalose, stress tolerance, geneflow.
Authors: Lars Holm Nelsen and Jørgen H. Seland
Language: Danish
Year: 2003
Vores formål med denne opgave har været at lave et program, som ud fra en beskrivelse af en række genstande i rummet kan lave en perspektivisk afbildning af disse genstande. Vores ambitioner har været at lave et program, der løser problemet på en enkelt og elegant måde, samt gennem intensiv afprøvning af programmet at dokumentere at programmet fungerer korrekt.
Den følgende rapport indeholder vores løsning på problemet. Vi bruger dog en anden metode, end den der er angivet i brevkassen til at oversætte planer i rummet til planer i skærmfladen. Dette skyldes at brevkassens metode først kom til efter, at vi havde designet og implementeret vores udgave.
Gennem den eksterne afprøvning har vi vist, at programmet overholder specifikationen på alle vigtige punkter. Vi har dog ikke formået at bringe alle fejlmeldninger op på et klart nok niveau.
Authors: Johan Jannik Bjerrum-Bohr, Tue Herlau and Lars Holm Nielsen.
Language: Danish
Year: 2003
Vores mål med denne opgave var at konstruere en strålesporer, der dels implementerede den funktionalitet, der var stillet i opgaven, dels implementerede forskellige ekstrafunktioner heriblandt antialiasing, reflektioner, flere lyskilder og først og fremmest let lod sig udvide med en stort vifte af ekstrafunktioner. Dernæst var et andet mål at implementere et såkaldt oktettræ for at gøre programmet hurtigere.
Vi mener i rapporten at have dokumenteret, at vi har opnået de strukturmæssige mål med programmet. Man kan let genbruge strålesporerdelen i andre sammenhænge. Funktioner så som et distribueret design, linseeffekter til kameraet, bedre lysmodel etc. kan implementeres med meget lokale ændringer i kildekoden.
Igennem afprøvningen mener vi ligeledes, at vi med den usikkerhed, der altid er hæftet til den slags, at have påvist at programmet kan håndtere de tegnemæssige krav, både de formelle og den ekstafunktionalitet, vi har tilført programmet.
Igennem vores afprøvning har vi ligeledes påvist, at oktetstrukturen fungerer tilfredsstillende, og vi har i vores udledninger af køretiden påvist, at den giver betydelige køretidsmæssige forbedringer. Vi har dog ikke haft tilstrækkelig mulighed til at afprøve, hvorledes disse køretidsmæssige forbedringer oversættes til faktisk tidsmæssige forbedringer ved faktiske strålesporinger.
Vores rapport og program er skrevet til personer med et kendskab til Java svarende til en dat0gb studerende eller højere.
Authors: Lars Holm Nielsen, Hans Christian Steen-Larsen, Kaspar Kirstein Nielsen and Jakob Molander
Language: Danish
Year: 2002
Vi har i dette projekt undersøgt Solens differentielle rotation udfra solpletter, samt rotationshastigheden ud fra solpletternes bevægelse henover solskiven. Dette har vi gjort ved at designe en række programmer, som først hentede Intensigram-billeder (taget med satellitten SOHOs MDI modul) fra perioden 1997-2003 fra SOHO's hjemmeside. Dette blev til 5800 billeder. Dernæst brugte vi programmerne til at bestemme solpletternes placeringer på billederne. Det sidste trin i data-processen, var at transformere plan-koordinaterne (rådata) om til heliografiske koordinater. Vi bestemte altså hver enkel solplets koordinater på deres vej hen over solskiven og derved Solens rotationshastighed. Vi valgte at dele solpletterne op i 4 kategorier: gruppe/single og gammel/ny. Resultaterne viser, at Solens rotationshastighed i allerhøjeste grad afhænger af, hvilken type solplet man ser på og hvilket tidspunkt i cyklussen solpletten tilhører. Endvidere har det vist sig, at Solens rotationshastighed ikke afhænger af solpletternes størrelse. For at kunne sammenligne med andres resultater, fitter vi til samme ligning som andre gennem det sidste århundrede har fittet til:
O(t) = A - B sin2(t)
hvor O er rotationshastigheden i grader/dag, A er rotationshastigheden i grader/dag ved ækvator, B er et mål for krumningen af kurven (grader/dag) og t er breddegraden. Med O forstås den synodiske rotationshastighed for Solen, da vi skal huske på, at billederne er hentet ned fra SOHO satelitten, der følger med Jorden rundt om Solen. Hvis ikke andet er nævnt er det den synodiske rotationshastighed vi hentyder til.
