Gravitating to rigidity: Patterns of schema evolution – and its absence – in the lives of tables
作者：
Highlights：
• We highlight patterns of table evolution for databases within open source software.
• Γ pattern: tables with large schemata tend to have long durations and survive.
• Comet pattern: the tables with most updates often are of medium schema size.
• Inverse Γ pattern: most tables change disproportionately lower wrt their duration.
• Empty Triangle pattern: deleted tables are mostly early born, short lived & quiet.
• Gravitation to rigidity: databases are much more prone to rigidity than evolution.
摘要
Highlights•We highlight patterns of table evolution for databases within open source software.•Γ pattern: tables with large schemata tend to have long durations and survive.•Comet pattern: the tables with most updates often are of medium schema size.•Inverse Γ pattern: most tables change disproportionately lower wrt their duration.•Empty Triangle pattern: deleted tables are mostly early born, short lived & quiet.•Gravitation to rigidity: databases are much more prone to rigidity than evolution.
论文关键词：Schema evolution,Database evolution,Analysis of evolution history,Patterns in schema evolution,Software rigidity,Software repository mining,Exploratory study,Software maintenance
论文评审过程：Received 26 March 2016, Accepted 30 June 2016, Available online 12 July 2016, Version of Record 10 September 2016.
论文官网地址：https://doi.org/10.1016/j.is.2016.06.010
