16th Working Conference on Reverse Engineering
- 14 Oct, 22:07 @akuhn: Nicolas Bettenburg uses Beanplots (rather than boxplots) to compare distributions of data, jstatsoft.org (via @leonmoonen #wcre09) ※ @
- 10:00 @onierstrasz: #wcre09 Static remodularization of software is analogous to web 1.0. Dynamic clustering is web 2.0. ※ @
- 21:53 @onierstrasz: #wcre09 anquetil: avoid "sparse" features to cluster (rarely occur, so does not help) ※ @
- 21:52 @onierstrasz: #wcre09 Instead of clustering to shoehorn software into new fixed structures, use it to generate dynamic views related to tasks at hand ※ @
- 21:52 @margaretstorey: Nice presentation by Anquetil/Lethbridge of influential paper from WCRE99 on clustering. Cluster using tags on work items? #wcre09 ※ @
- 21:52 @onierstrasz: #wcre09 anquetil 1999 paper: choice of clustering algorithm does not matter (!); choice of similarity metric does not matter (much) ※ @
- 21:51 @onierstrasz: #wcre09 most influential paper 1999- Nicholas Anquetil : clustering algorithms do not discover hidden structure in software; they impose it ※ @
- 9:30 @onierstrasz: #wcre anquetil: avoid "sparse" features to cluster (rarely occur, so does not help) ※ @
- 9:26 @onierstrasz: #wcre most influential paper 1999- Nicholas Anquetil : clustering algorithms do not discover hidden structure in software; they impose it ※ @
Generated by Twit2Html on 6 November 2009.