André > Recent Status Updates

The proposed Evolution Radar is not a bad coupling visualization idea, but the overplotting problem of graph-based visualization can also be solved by the fact that each graph can be represented as a matrix. So a numerical Design/Dependency Structure Matrix (DSM) would be an alternative representation that has already proven itself for coupling visualization (as the name suggests) and is less proprietary (2010 book)

"... while researchers considered the presence of change coupling a bad symptom, nobody empirically assessed whether it correlates with the presence of software defects"

D'Ambros mentions the mining of softw.-repos for change-coupling data to inform re-designs, ie, what changes together logically belongs together (as opposed to a naive functional system decomposition, cf Juval Löwy's volatility-based decomposition)

Türkei, Griechenland und Vertrag von Lausanne (1923) wird wohl wieder häufiger Thema...

- Regelkonformitaet≠Wahrheit
- rein log. Aussagen sind Tautologien (keine neue Erkenntnis)
- bequem≠genau
- Mathe≠formale Logik
- Geltung math. Erkenntnisse nicht gleichgueltig ggue _Bedeutung_ von Symbolen
- SoWi-Mathematisierung abstrahiere von Ergebnissen der Mathe; Mathe sei nicht bloss Erkenntnis ihrer Gegenstaende, sondern Loesung nichtmath., epistemischer Probleme
- Nutzentheorie als unwissenschaftl. Mathemat.