"Another issue is non-destructively replaying development changes into test/production environments."
"...you need to maintain referential integrity when replaying those updates/inserts"
I think most of your misgivings about this tool are grounded in the expectation that it would provide merge features like Git. Since it claims to be "git for databases", I suppose that's a fair expectation.
However, the tool does provide the core functionality of Git, which is to implement a tracking system for hash-tagged objects (or, in this case, tables). This does have value. Not only is it faster than standard db dumps, it's also more space-efficient. For some people, this could be really valuable.
I think it's an interesting tool and could serve as a launching point for more powerful/useful functionality.
"...you need to maintain referential integrity when replaying those updates/inserts"
I think most of your misgivings about this tool are grounded in the expectation that it would provide merge features like Git. Since it claims to be "git for databases", I suppose that's a fair expectation.
However, the tool does provide the core functionality of Git, which is to implement a tracking system for hash-tagged objects (or, in this case, tables). This does have value. Not only is it faster than standard db dumps, it's also more space-efficient. For some people, this could be really valuable.
I think it's an interesting tool and could serve as a launching point for more powerful/useful functionality.