The technical structure of the Mexican science and technology literature was determined. A representative database of technical articles was extracted from the Science Citation Index for the year 2002, with each article containing at least one author with a Mexican address. Many different manual and statistical clustering methods were used to identify the structure of the technical literature (especially the science and technology core competencies), and to evaluate the strengths and weaknesses of each technique. Each method is summarized, and its results presented.

The similarity of documents in a large database of published Fractals articles was examined for redundancy. Three different text matching techniques were used on published Abstracts to identify redundancy candidates, and predictions were verified by reading full text versions of the redundancy candidate articles. A small fraction of the total articles in the database was judged to be redundant. This was viewed as a lower limit, because it excluded cases where the concepts remained the same, but the text was altered substantially.Far more pervasive than redundant publications were publications that did not violate the letter of redundancy but rather violated the spirit of redundancy. There appeared to be widespread publication maximization strategies. Studies that resulted in one comprehensive paper decades ago now result in multiple papers that focus on one major problem, but are differentiated by parameter ranges, or other stratifying variables. This paper inflation' is due in large part to the increasing use of metrics (publications, patents, citations, etc) to evaluate research performance, and the researchers' motivation to maximize the metrics.

: Database Tomography (DT) is a textual database analysis system consisting of two major components: 1) algorithms for extracting multi-word phrase frequencies and phrase proximities (physical closeness of the multi-word technical phrases) from any type of large textual database, to augment 2) interpretative capabilities of the expert human analyst. DT was used to obtain technical intelligence from a Wireless LAN (Local Area Network) database derived from the Science Citation Index/ Social Science Citation Index (SCI). Phrase frequency analysis by the technical domain experts provided the pervasive technical themes of the Wireless LAN database, and the phrase proximity analysis provided the relationships among the pervasive technical themes. Bibliometric analysis of the Wireless LAN literature supplemented the DT results with author/journal/institution publication and citation data.

Text mining was used to derive technical intelligence from an abrupt-wing-stall database derived from the Science Citation Index database. Both concept and document clustering were used to provide the structural taxonomy of the global abrupt-wing-stall literature and the estimated relative levels of effort in the major subcategories. Bibliometric analysis of the abrupt-wing-stall literature generated author/journal/institution publication and citation data.

This paper describes Paradis-Net, a typed event-driven message-passing interface for designing distributed systems. Paradis-Net facilitates the development of both peer-to-peer and client-server architectures through a mechanism called "Cooperation". We introduce the programming interface and compare its mechanisms to active messages and remote procedure calls. Finally we demonstrate how the interface can be used to implement communication patterns typical for distributed systems and how peer-to-peer functionality can be mapped onto ParadisNet.

Im Sommersemester 2004 wurde im Seminar Grid Computing und Peer-to-Peer eine Reihe aktueller Themen aus den Grid Computing, Peer-to-Peer Systeme und Ad-Hoc Netzwerkem Gebieten angeboten. Jeder Teilnehmer wahlte hieraus ein Thema, um daruber in der Form eines medial gestutzten Vortrages zu referieren. Um allen Teilnehmern die Gelegenheit zu geben, aus diesem Seminar nachhaltig etwas mitzunehmen, fertigte jeder Vortragende eine allen zugangliche schriftliche Ausarbeitung an. Die Ausarbeitungen finden sich in leicht redigierter Fassung durch die Editoren im vorliegenden technischen Bericht wieder.

This paper describes Paradis-Net, a typed event-driven message-passing interface for designing distributed systems. Paradis-Net facilitates the development of both peer-to-peer and client-server architectures through a mechanism called “Cooperation”. We introduce the programming interface and compare its mechanisms to active messages and remote procedure calls. Finally we demonstrate how the interface can be used to implement communication patterns typical for distributed systems and how peer-to-peer functionality can be mapped onto Paradis-Net.

The structure and infrastructure of the Mexican technical literature was determined. A representative database of technical articles was extracted from the Science Citation Index for the year 2002, with each article containing at least one author with a Mexican address. Many different manual and statistical clustering methods were used to identify the structure of the technical literature (especially the science and technology core competencies). One of the pervasive technical topics identified from the clustering, thin films research, was analyzed further using bibliometrics, in order to identify the infrastructure of this technology.

Database Tomography (DT) is a textual database analysis system consisting of two major components: (1) algorithms for extracting multi-word phrase frequencies and phrase proximities (physical closeness of the multi-word technical phrases) from any type of large textual database, to augment (2) interpretative capabilities of the expert human analyst. DT was used to obtain technical intelligence from a Fractals database derived from the Science Citation Index/Social Science Citation Index (SCI). Phrase frequency analysis by the technical domain experts provided the pervasive technical themes of the Fractals database, and the phrase proximity analysis provided the relationships among the pervasive technical themes. Bibliometric analysis of the Fractals literature supplemented the DT results with author/journal/institution publication and citation data.

Since the rise of the light weight software processes, the paradigm on how software should be developed has started to shift. Agile methods strive to supersede the traditional software process with its exhausting requirements elicitation at the beginning of a software development project, at least for smaller or younger companies. The software engineering group at the Universität Karlsruhe has accounted for this shift and extended their offer of lectures by an Extreme Programming lab course held in cooperation with andrena objects ag.

This paper presents the integration of two collective I/O techniques into the Clusterfile parallel file system: disk-directed I/O and two-phase I/O. We show that global cooperative cache management improves the collective I/O performance. The solution focuses on integrating disk parallelism with other types of parallelism: memory (by buffering and caching on several nodes), network (by parallel I/O scheduling strategies) and processors (by redistributing the I/O related computation over several nodes). The performance results show considerable throughput increases over ROMIO's extended two-phase I/O.

JPlag is a web service that finds pairs of similar programs among a given set of programs. It has successfully been used in practice for detecting plagiarisms among student Java program submissions. Support for the languages C, C++ and Scheme is also available. We describe JPlag's architecture and its comparsion algorithm, which is based on a known one called Greedy String Tiling. Then, the contribution of this paper is threefold: First, an evaluation of JPlag's performance on several rather different sets of Java programs shows that JPlag is very hard to deceive. More than 90 percent of the 77 plagiarisms within our various benchmark program sets are reliably detected and a majority of the others at least raise suspicion. The run time is just a few seconds for submissions of 100 programs of several hundred lines each. Second, a parameter study shows that the approach is fairly robust with respect to its configuration parameters. Third, we study the kinds of attempts used for disguising plagiarisms, their frequency, and their success.

Finding changed identifiers is important for understanding the difference between two versions of a program and for detecting and resolving conflicts while merging variants of a program together. Standard practice for differencing and merging relies on line based techniques that do not recognize renamed identifiers. The design and implementation of a tool to automatically detect renamed identifiers between two versions of a program is presented. The system uses an abstract representation of language constructs to enable language awareness without introducing language dependence. Modules for Java and Scheme have been written. The detector works with multiple file pairs, taking into account renamings that span several files. A case study is presented that demonstrates proof of concept. The detector is part of a suite of intelligent differencing and merging programs that exploit the static semantics of programming languages.