In workshop we can call action types only on user click (button, metric card) or indirectly through automations.
We have cases where we would like action types to be called based on other triggers like e.g. variable changes without the user having the need to click a button.
Since this is not a variable per se I would see this as a new category like variables, layouts → „triggered action types calls“. It is basically a light form of automations - but with local variables within a workshop module. The set up of each triggered call I also imagine very similar to automations (but less options ). Basically selecting the trigger, optional conditions (like comparison vs value thresholds) and wiring additional input variables to the call.
Triggers I can think of:
on variable change (with optional conditional check)
timed?
A couple examples where we would use this:
numeric/text inputs automatically update objects without the user having the need to click a submit button
use parent/loop interface variable to trigger the same action type call in all loops at the same time.
usage telemetry like tracking viewed sections in a workshop module (is visible bool used for collapsible sections)
This has been discussed before, and the conclusion was that this would be dangerous.
A user could misconfigure a function to e.g. delete every object in an ObjectSet.
A lot has changed since then, and I could also see many cases where this would be useful. Being able to revert has made the cost of errors lower for developers. And there could be a flag to toggle, where the user accepts the risk of editing the Ontology based on an event trigger.
This is one of the reasons why we’ve built several workflows with OSDK instead. Workshop either didn’t allow for the proper functionality – relatively small things making workflows feel unintuitive or cumbersome – or the UI was not customisable enough, making the application look cheap or clunky.
Thanks @jakehop! Indeed it holds a risk to allow event triggered action type calls. However, isn’t it the same/similar risk on automations or user executed calls? We as builders always need to ensure that we don’t miss-configure an automation or wire the wrong object set into a user-executed button click
). Basically selecting the trigger, optional conditions (like comparison vs value thresholds) and wiring additional input variables to the call.