Writings

fileno(3) on C++ Streams: A Hacker's Lament
on the web
Abstract: There is a need to bridge the gap between high-level programming concepts and low-level operating system facilities. On Unix, we occasionally need to directly access the file descriptor pertaining to a C++ I/O stream. This, however, is much more complicated than it should be. I try to provide a partial solution.

with Isabella Bierenbaum and Dirk Kreimer:
On the Invariance of Residues of Feynman Graphs
local PS; MZ-TH/01-33; hep-th/0111192; DOI 10.1063/1.1502926; J.Math.Phys. 43, 4721-4740
Abstract: We use simple iterated one-loop graphs in massless Yukawa theory and QED to pose the following question: what are the symmetries of the residues of a graph under a permutation of places to insert subdivergences. The investigation confirms partial invariance of the residue under such permutations: the highest weight transcendental is invariant under such a permutation. For QED this result is gauge invariant, ie the permutation invariance holds for any gauge. Computations are done making use of the Hopf algebra structure of graphs and employing GiNaC to automate the calculations.

with Christian Bauer and Alexander Frink:
GiNaC
local PS; Contribution to the Handbook of Computer Algebra edited by Johannes Grabmeier, Erich Kaltofen and Volker Weispfenning
This paper is a very short description of the GiNaC framework for symbolic computation within the C++ programming language.

with Christian Bauer and Alexander Frink:
Introduction to the GiNaC Framework for Symbolic Computation within the C++ Programming Language
local PS; MZ-TH/00-17; cs.SC/0004015; DOI 10.1006/jsco.2001.0494; J. Symbolic Computation, 33, 1-12
Abstract: The traditional split-up into a low level language and a high level language in the design of computer algebra systems may become obsolete with the advent of more versatile computer languages. We describe GiNaC, a special-purpose system that deliberately denies the need for such a distinction. It is entirely written in C++ and the user can interact with it directly in that language. It was designed to provide efficient handling of multivariate polynomials, algebras and special functions that are needed for loop calculations in theoretical quantum field theory. It also bears some potential to become a more general purpose symbolic package.

Parallelization of adaptive MC Integrators - Recent pvegas developments
local PS; MZ-TH/98-54; physics/9812011
Abstract: This paper shortly describes some important changes to the pvegas-code since its first publication. It proceeds with a report on the scaling-behavior that was found on a wide range of current parallel hardware and diScussesscusses some issues of optimization that may be thrown up.

with Dirk Kreimer and Karl Schilcher:
First Results with a new Method for calculating 2-loop Box-Functions
MZ-TH/98-16; hep-ph/9804333; DOI 10.1007/s100529800947; Eur.Phys.J. C, 4, 693-699
Abstract: We describe a first attempt to calculate scalar 2-loop box-functions with arbitrary internal masses, applying a novel method proposed in [1]. Four of the eight integrals are accessible to integration by means of the residue theorem, leaving a rational function in the remaining variables. The result of the procedure is a three- or sometimes two-dimensional integral representation over a finite volume that can be further evaluated using numerical methods.

Parallelization of adaptive MC Integrators
MZ-TH/97-30; physics/9710028; Comp. Phys. Comm., 106, 258-266
Abstract: Monte Carlo (MC) methods for numerical integration seem to be embarassingly parallel on first sight. When adaptive schemes are applied in order to enhance convergence however, the seemingly most natural way of replicating the whole job on each processor can potentially ruin the adaptive behaviour. Using the popular VEGAS-Algorithm as an example an economic method of semi-micro parallelization with variable grain-size is presented and contrasted with another straightforward approach of macro-parallelization. A portable implementation of this semi-micro parallelization is used in the xloops-project and is made publicly available.

„Gut gemeint“ und „gut gemacht“…
UNIPRESS 309, 1997; im Web
Was passieren kann, wenn arglos Informationen ins Web gestellt werden, die niemals in die Hände von Außenstehenden gelangen sollten. Eine Mahnung, in der einige vernachlässigte Risken in der aufkommenden Ära der Suchmaschinen erörtert werden.

Buchbesprechung: Unbestimmtheit oder Krieg?
Physilis 40, 1996; im Web
Eine Doppelbuchbesprechung zweier neuerer Heisenberg-Biographien: Thomas Powers Heisenberg's War und David Cassidys Uncertainty.

mit Volker Schneider:
Beschreiben statt bewerten. Quantitative deskriptive Sensorik deutscher Rieslinge
Das Deutsche Weinmagazin 9; 1995, 15-24
Die praktische Sensorik beschränkt sich in Deutschland weitgehend auf die Bewertung der Qualität. Geruchs- und Geschmacksmerkmale werden von den Bewertungsschemen nicht erfaßt. Um diese zu erfassen, bedient man sich in anderen Ländern schon seit längerem der "quantitativen deskriptiven Sensorik". Mit diesem Verfahren wurden nun erstmals 15 deutsche Rieslingweine verglichen.