Open christian-konrad opened 2 years ago
I've argued for "not being special" here: https://github.com/camunda/camunda-modeler/issues/2883#issuecomment-1088512467
My believe is our energy is invested better, elsewhere :slightly_smiling_face:.
As a counterargument tracking application clicks of buttons and links is now just a matter of configuration and requires very little dev work. So I wouldn't mind setting it up if it is useful to determine if these items should stay. I absolutely agree that not being special is the way to go, but it's even better if we can support that with data.
Although that only stands for click events, if we want more fancy tracking then it is work :P
I have added a comment in the referenced issue.
Click tracking is definitively a helpful data point. We should not only track the event itself (e.g. close all tabs) but also the UI element (e.g. right click menu, context menu, shortcut).
As a counterargument tracking application clicks of buttons and links is now just a matter of configuration and requires very little dev work.
@Skaiir How is that "very little dev work" without https://github.com/bpmn-io/internal-docs/issues/507? Please comment on that issue if we have everything in place already (be able to track DOM interactions with context, easily, and fail-safe).
@nikku We already track all application clicks across the whole modeler, and we currently filter based on two conditions: is it a child of parent with ID x (providing context) && is it a button / anchor. We can add new parents and automatically all the buttons and links within them will be tracked. It's primitive but it works for quick datapoints.
@Skaiir Please link to our existing solution.
Definitely not the most fancy or efficient of implementations but very noninvasive. I extended this a little when requirements asked for the welcome page to be tracked as well. Adding tracking for a section of the app would involve adding an HTML ID and adding that value to the collection at the top of the referenced code file.
Thanks. I think that goes into a good direction and can be built upon.
Is your feature request related to a problem? Please describe.
There are users not understanding the current behavior of tab management: https://github.com/camunda/camunda-modeler/issues/2883 It will be more of a concern once we support multi-file projects (https://github.com/camunda/product-hub/issues/59).
Describe the solution you'd like
To support the understanding of current and future user behavior for tab management in a multi-file project context, we want to track the usage of those actions.
Describe alternatives you've considered
Alternatives would be usability tests, guerilla tests in-house or just doing nothing here, but following best practices of popular IDEs.