You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I think so far we're mainly focusing on "pull-pull-push one-way sync", right?
Read from the source, read from the target, calculate which changes are necessary and write them to the target.
My original idea was a pull-pull-push one directional that can also work in the opposite direction. That is, when pulling, you can get the "offset of changes" on both of the sources and compare them. Meaning it can as easily work for a two-way sync. The assumption here is that the sync will assume everything is in the intended state at the time that the sync is switched on.
This approach covers all the paradigms except the federated query.
It would contradict the conversation we had earlier about the change interpreter describing the desired state. To be honest, the difference between these paradigms were not as clear in my mind when I came up with this model.
Still, both reader and writer code interfaces are difficult to come up with. Imagine one task tracking reader mentions a project A which the writer for the target app will have to make sure such a project is present or define it otherwise.
From federatedbookkeeping/research#36 (comment) -
CYB should probably support multiple sync paradigms:
The text was updated successfully, but these errors were encountered: