Regarding Vaadin 8, 10, etc.
Sorry if I wasn’t clear enough. Of course I didn’t mean that we are going to stay with Vaadin 8 forever. We successfully migrated from Vaadin 5 to 6 and then to 7 a while ago, and through all these migrations our GUI API and XML schema remained relatively stable, at least comparing to changes in Vaadin. Now we are working on the migration to Vaadin 8, and again we take all the heavy lifting on ourselves. For example, the data binding in Vaadin 8 has changed dramatically, it now provides only one-way bindings. So we are now implementing the whole data binding mechanism on our side for to keep your code compatible.
We will start thinking about further migrations as soon as we complete the current one, probably next year (2019). Will it be Vaadin 10, Vaadin 20, a fork or a hypothetical homegrown framework - I simply don’t know at the moment. Anyway, we understand the importance of the compatibility for long-living enterprise projects, so it is always our priority.
Regarding the BPM addon. We are now evaluating two ways of improving it:
Continue with the current implementation based on Activity. Upgrade to the latest version and put efforts into improving the API and usability.
Switch to Camunda or Flowable engines and modelers. Make significant changes in the integration architecture if we decide that it is needed.
Both ways assume that the current processes will continue to run. In the second approach, a new BPM-2 addon will appear, but the current one will be supported too, just without further active development. So I wouldn’t recommend you waiting for the new implementations if you are already using the BPM addon or planning to employ it in the near future. All your processes and business logic will continue working as long as they use the current public API, possibly requiring minor migrations. However, I’m not sure about your customizations, it depends on what you are doing.