André > Recent Status Updates

Confusingly, author treats "quality" as a peer to other ilities (safety, ...) rather than a umbrella term defined through ilities (safety, maintainability, flexibility, ...). He even constructed an ilities correlation network from thousands of websites and—surprise, surprise—found 'quality' at the center. Category mistake already by his own def of quality as "well made"/low tolerances (traits of ilities themselves)

Einführung, Begriffsunterscheidungen (US-Maintenance vs DE-Wartung, Erhaltung vs Entwicklung, Evolution usw), Wirtschaftlichkeit: Werterhaltung von Sw-Kapitalgütern nach Hayek, Lachmann ua, wertgetriebene Sw-Evolution, Einflüsse auf Evolutionskosten, Kosten-Nutzen-Schätzung

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)