Ladies and gentlemen, i am disappoint.
I merely tried to install galaxy to another machine and test if our workflow gives the same result as on the old machine. And what i see?
First, i exported the workflow and imported it into the new galaxy, and despite both galaxies being one exact release, i got an error. Workflow submission failed Errors happen, i know, and it would be ok, but (a) the error text gave almost no clue what gone wrong and where, so i had to investigate exact problematic location myself (within 55 steps of workflow), and (b) even with professional help i could not get reliable solution.
Second. In order to workaround mentioned error i had to save a copy of the workflow in the workflow editor. And as i did it, i saw that numbers of steps changed despite i did nothing in the editor but saved a copy. It would also be ok if i could identify the steps in some other way, but, as we soon will see, i could not.
Third. Step numbers in the history do not correspond to step numbers in the workflow invoked to make this history.
Fourth. In the history entry there is no place for original step name. E. g. if my step was called “MergeBamAlignment normal exome all reads and mapped reads”, in the history it magically turns into “MergeBamAlignment on data 1, data 2, and data 7: BAM with merged alignments” ! Or data 3, data 5, and data 7, i am not sure. And i found no field containing the original name.
Fifth. These data numbers are the same in the input and output history, so “data 1” may mean my first uploaded file or result of the first processing step.
Ok, you might say, you can “view details” and see the exact inputs of the steps of interest and even its parameters. Sure i can. But why make users conduct their own investigation like sherlock holmes every time they need to know what step they are looking on, while you can simply call history item “MergeBamAlignment normal exome all reads and mapped reads on data i1, data i2, and data o7” instead of “MergeBamAlignment on data 1, data 2, and data 7”? I heard galaxy was made to offer better traceability than the good old command line and bash/python scripts. But i can not see this traceability yet.
Thank you for your attention.