Suggested experiment setups provided by the benchmark

[denkv]

User story

Benchmark authors may want to define multiple experiment setups (parameter values) which can be easily run.

Solution

Define a vocabulary for describing experiment setups in benchmark.ttl and provide a list of predefined experiments in UI. One should be able to change parameters before actually running it. Being able to run several different experiments from that list (and all of them) will be helpful too.

See also https://github.com/hobbit-project/platform/issues/448

• [ ] define vocabulary for defining experiment setups and for linking them to a benchmark resource
• [ ] list experiment setups in UI and allow to use them to pre-fill the form then edit it if desired

Another possibility would be to define sensible ranges of parameters and a button to randomize them during the experiment setup.

See also

There already is http://w3id.org/hobbit/vocab#defaultValue for specifying the default parameter values, but this approach does not allow to easily specify several sets of parameter values.