The myExperiment website was launched in November 2007 and contains a significant collection of scientific workflows for a variety of workflow systems, most notably Taverna, but also other tools such as Bioclipse. myExperiment has a REST API and is based on an open source Ruby on Rails codebase. It supports Linked data and had a SPARQL Endpoint, with an interactive tutorial.[6][7][8][9]
While the initial idea of myExperiment may have been to capture computational experiments as executions of registered workflows (submitted to a configured workflow runner API), in reality most users of myExperiment shared only workflow definitions, possibly with example files.
Influence on Research Objects development
The myExperiment Ontology[10] included details of social networking, attributions, experiments and workflow structures. This work was instrumental in forming the idea of Research Objects,[11] a Linked Data method for identification, aggregation and exchange of scholarly information. In the myExperiment Ontology, the Research Object was loosely represented as a "Pack" of typed related resources, including workflows.
myExperiment was since enhanced by the Workflows Forever project (Wf4Ever)[12] which in 2011 aimed to provide new features to support the preservation of Research Objects in conjunction with the dLibra digital library framework.[13][14]
The Wf4Ever co-development with the workflow system Taverna (at the time maintained by myGrid), led to the creation of the Research Object Model ontologies.[15] RO Model uses many of the conceptual ideas of the myExperiment Ontology, but with a basis of an OAI-ORE aggregation as the central digital object, and with a primary goal of reproducibility and integration with the W3C provenance model PROV, the replacing RO Model was independent of the social network at a workflow registry.
In 2019, subsequent development of Research Object Crate (RO-Crate) replaced the RO Model with a profile of schema.org serialised in JSON-LD, this was general purpose for scholarly outputs. This ontology-less approach was better suitable for packaging in a ZIP archive and aimed to be more approachable for developers without Linked Data experience.[16]
The research group has since launched[18][19] myExperiment's replacement, the workflow registry workflowhub.eu,[20] which uses a profile of RO-Crate for publishing workflows together with their context and dependencies. In contrast with myExperiment, the WorkflowHub however do not focus on the social network aspects (beyond institutional/project organisation of entries), and do not attempt to capture any experiments.
^Gewin, Virginia (2008). "The new networking nexus: A crop of websites is making networking among scientists easier than ever". Nature. 451 (7181): 1024–1025. doi:10.1038/nj7181-1024a. PMID18363200.