We question whether analytical tools such as Common Workflow Language, which aim to make computational methods “reproducible and shareable”, can stand the test of time (see Nature 573, 149–150; 2019).