data replication

[ghost]

clocker's internals could be reworked to write operations into an append-only log with hyperlog so that the clocker database can easily be replicated among different machines or across teams of people. The existing command-line interfaces can largely be preserved with a few small modifications to deal with occasional multi-value records. The list and data commands can read from a materialized view of the data created by a hyperlog index.

There was some discussion of this feature already in #12. If nothing else, this would be a very useful backup feature, and this indexing scheme will make it easier to generate different views of the underlying data in the future. This feature should also be integrated with a migration script to automatically port existing data into the new format.

[fnogatz]

I would prefer to first discuss a list of goals we want to achieve with clocker. Currently I like it especially for its CLI and simplicity. For my personal use I only need a way to work across multiple machines, either by a git-like pull/push workflow or by a silently shared database. Plus even better CLI interfaces to pipe clocker's output to tools like cli-gh-cal. I would expect that teams would prefer some other time tracking tool.

[fnogatz]

ping @substack