Documentation should say explicitly that costs in #maximize are turned negative by their compilation to #minimize statements.

[chenkins]

From a modeller's perspective, who sees #maximize, this seems counter-intuitive and should be made explicit.

The following section should be expanded:

Note that, with #maximize statements in the input, this correlation might be less straightforward because they are compiled into #minimize statements in the process of generating smodels format [80].

Add something like

In the solver output, costs which are maximized are output as their negative counterpart (which is minimized). Example....

[tortinator]

Merci vielmals, @chenkins!