Author: Jack Maguire, Created January 2018

Checklist of StandardJobQueen virtual functions that you generally want to override



The goal of this tutorial is to help developers get experience with writing their own JD3 protocol. You can use this page for reference to JD3 concepts throughout the walkthrough. I will link to specific sections of that page when they become relevant, so there is no need to read the whole thing now. That said, reading this section would be a good place to start.

For the sake of having an example to work on, let's pretend that we want to have a protocol where mover A and mover B are both run on the input structure and the results are merged and run through mover C. See step 2 if it is not clear what I mean

In general, the code generated in this example is NOT written with effeciency, performance, or even our own coding conventions in mind. The main goal is clarity and simplicity/readability. All feedback is welcome! Please send thoughts/questions to

Changes in the code since this tutorial was written

JobGenealogist no longer takes local_job_id

While I was writing this tutorial, I was frustrated by how the JobGenealogist would sometimes take the global_job_id and other times the local_job_id. It is very easy to pass the wrong element and not notice. So I changed the JobGenealogist such that it only ever interacts with the global_job_id. I tried to go through the tutorial and find all the things that needed to be updated, but it is very possible I missed a few. Please keep this change in mind if you see code that does not quite make sense.



Beside each link is a list of the virtual function overrides that we introduce in that section.


Step 1: Skeletons for creating the file

Step 2: Creating the Job DAG initial_job_dag() parse_job_definition_tags()

Step 3: Node Managers

Spawning Jobs

Step 4: Creating Larval Jobs determine_job_list()

Step 5: Writing a Job Class

Step 6: Maturing Larval Jobs complete_larval_job_maturation()

Processing Results

Step 7: Note Job Completed note_job_completed()

Step 8: Completed Job Summary completed_job_summary()

Step 9: Discarding Job Results job_results_that_should_be_discarded()

Step 10: Outputting Results jobs_that_should_be_output()

Bonus Content


Appendix A: Let the user define their own score function in <Common> append_common_tag_subelements()

Appendix B: Let the user define individual score functions in <Job> append_job_tag_subelements()

Helpful Classes

Node Manager

Job Genealogist

See Also