I see what others is suggesting however, I don't see anything wrong with figure 1, the outputs to me are vary logical in the way they are shown.
The top side PNP is being biased in current mode which is simply supplying the upper NPN ? Actually, I kind of like the way it's laid out. Of Course voltage output swing specs are suggesting a conventional darlington configuration but I don't think this is the case here.
The bottom side is kind of cool too. That type of configuration is much like a bakers clamp idea to prevent over charge however, it does exhibit the problem of not taking advantage of a very low voltage swing on the low rail.
This chip looks like a perfect candidate for a project we have brewing to drive a fuel injector. The injectors we are experimenting with uses ~ 600 ma to open. They are little inductive and we need to use a wheeling diode back to the (+) rail to suppress it. I noticed the diode in that chip is also in the exact position where I would need it however, I don't see any comments on the sheet about current handling capabilities?
Jamie