Workflow process modeling alternatives?

[Timothy Fallon]

Hi there all,

I see PAML is (somehow) built on UML. My first introduction to these types of languages(?) - or process description standards I suppose is the better terminology, was BPMN. I am not directly familiar with UML. But I briefly came across YAWL, as described on http://www.workflowpatterns.com/ , https://yawlfoundation.github.io/ , https://www.yaug.org/ , & it seemed like it was a bit more "bulletproof" than BPMN for building something like PAML upon it. I haven't looked into the specifics of YAWL vs UML, but I'm curious if YAWL was evaluated & what the relevant pros/cons might be vs UML? This is just for my own curiosity. I'm not suggesting PAML should switch to YAWL.

Admittedly, the symbology of UML, seems a bit clearer that YAWL (as detailed in Fig 1 of this PDF: https://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.159.2913&rep=rep1&type=pdf), but the PAML editor seems to be using its own symbology / graphics vs the symbology of UML?

Here is another recent paper which I found helpful for thinking about YAWL. If there are similar resources for UML, please let me know. Would be happy to read up!

Adams M, Hense AV, ter Hofstede AHM. 2020. YAWL: An open source Business Process Management System from science for science. SoftwareX 12:100576. doi:10.1016/j.softx.2020.100576

All the best,

-Tim

[Jake Beal]

Hi, Tim:

Thank you for sharing these references. While I was not previously aware of YAWL,from what you shared it looks like another close cousin from the workflow language family. As all of these languages are derived from Petri nets, it is not surprising that they have some strong similarities. The key thing that we were looking for in the choice of a workflow language for PAML was that it be 1) universally expressive, including for distributed asynchronous execution, and 2) well established with strong support over a long history of use.  UML fit that bill well, and it's not surprising that there are other alternatives that could be competitive with it.

Thanks,

-Jake

--
You received this message because you are subscribed to the Google Groups "bioprotocols" group.
To unsubscribe from this group and stop receiving emails from it, send an email to bioprotocols...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/bioprotocols/df89490d-b837-4450-bad6-7e0eae563fc7n%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Timothy Fallon]

Hi there Jake,

Thanks for the feedback! Makes sense to me.

BPMN checks off the (2) requirement, it has been used for bio-process type protocol description. Here are two interesting examples I know of:

- Unitelabs built their prototype(?) Uniteflow workcell scheduler using BPMN (I believe the scheduler is an adaptation of the Camunda BPMN executor)

-  Roche has used Camunda/BPMN internally. See video: https://www.youtube.com/watch?v=B-LU1Tq7qYw 

I don't know if BPMN (or YAWL) would check off the (1) requirement. My expectation is that YAWL would check off (1), as it was designed by academics over a decade or so to capture as much of the possible workflow description logic patterns that had been seen in the field (enumerated in http://www.workflowpatterns.com/patterns/control/ ,http://www.workflowpatterns.com/patterns/resource/ , etc. , see website). The closest thing I know to a Petri net is probably a fungus or Streptomyces that grows on agar ;)

I guess a follow up question would be: What are the advantages of a workflow description language vs a general language (e.g. Scala like is used by Radix for their platform, or Python examples like the LARA suite , or the Opentrons OT2 API). Feel free to point me to a resource to read if these trade-offs have already been discussed! Or if they are planned to be discussed in an upcoming manuscript, happy to wait until that might be available.

All the best,

-Tim

[Jake Beal]

Hi, Tim:

There is an article on the design of PAML that we submitted in December. I'm attaching it, but I'm also thinking maybe we should get it up on a preprint server...

Thanks,

-Jake

To view this discussion on the web visit https://groups.google.com/d/msgid/bioprotocols/174afcdb-5321-4097-8d2a-a1f9fd492207n%40googlegroups.com.

[Ben Keller]

I did look at the business workflow languages and they did not seem like a good fit in the context of how lab procedures are modeled in Aquarium and the systems for biofabs that were involved in SD2.

Ben

---

From: biopro...@googlegroups.com <biopro...@googlegroups.com> on behalf of Jake Beal <jake...@gmail.com>
Sent: Wednesday, February 2, 2022 9:35 AM
To: Timothy Fallon <tfa...@ucsd.edu>
Cc: bioprotocols <biopro...@googlegroups.com>
Subject: Re: Workflow process modeling alternatives?

 

To view this discussion on the web visit https://groups.google.com/d/msgid/bioprotocols/CALR99W15WXXJ31dHxg69mmmSHDfvRgCh2VwR_kQ7GpOnih9YxQ%40mail.gmail.com.

[Timothy Fallon]

Thank you Jake & Ben for the feedback! Interesting & informative to read. I will also note: some interesting technical details on the of UML in PAML have been discussed in a few different places in the PAML Github issues (I found recently via searching), & those were interesting to read as well, although I didn't appreciate them fully.

On the manuscript, not sure if you are still taking comments, but I have some brief ideas for citations that could be added (citations that I found especially helpful as learning more about this workflow description field as a wet & dry lab biologist - unrelated to my current work). I can direct message those comments if interested.

All the best,
-Tim

[Jake Beal]

Please do send suggestions and comments by DM --- the manuscript is under review and will have opportunities for improvement.

Thanks,

-Jake

To view this discussion on the web visit https://groups.google.com/d/msgid/bioprotocols/65171cd5-b441-4adb-9330-4b9827aeffean%40googlegroups.com.