Errors in setting up kinds and how they multiply are sometimes caught by the kinds module: this section collects and issues such errors as tidy Inform problems.

§1. The :kinds group of tests for inform7 generates all of the problems below.

define PROBLEM_KINDS_CALLBACK KindsProblems::kinds_problem_handler

void KindsProblems::kinds_problem_handler(int err_no, parse_node *pn, text_stream *E,
    kind *K1, kind *K2) {
    switch (err_no) {
        case DimensionRedundant_KINDERROR:
            StandardProblems::sentence_problem(Task::syntax_tree(), _p_(PM_DimensionRedundant),
                "multiplication rules can only be given once",
                "and this combination is already established.");
            break;
        case DimensionNotBaseKOV_KINDERROR:
            StandardProblems::sentence_problem(Task::syntax_tree(), _p_(PM_DimensionNotBaseKOV),
                "multiplication rules can only involve simple kinds of value",
                "rather than complicated ones such as lists of other values.");
            break;
        case ImproperSubtraction_KINDERROR:
            StandardProblems::sentence_problem(Task::syntax_tree(), _p_(PM_ImproperSubtractionKOV),
                "we can only specify the result of one kind minus itself",
                "and moreover it has to be a dimensionless kind.");
            break;
        case NonDimensional_KINDERROR:
            StandardProblems::sentence_problem(Task::syntax_tree(), _p_(PM_NonDimensional),
                "multiplication rules can only be given between kinds of "
                "value which are known to be numerical",
                "and not all of these are. Saying something like 'Pressure is a "
                "kind of value.' is not enough - you may think 'pressure' ought "
                "to be numerical, but Inform doesn't know that yet. You need "
                "to add something like '100 Pa specifies a pressure.' before "
                "Inform will realise.");
            break;
        case UnitSequenceOverflow_KINDERROR:
            StandardProblems::sentence_problem(Task::syntax_tree(), _p_(PM_UnitSequenceOverflow),
                "reading that sentence led me into calculating such a complicated "
                "kind of value that I ran out of memory",
                "which my programmer really didn't expect to happen. I think you "
                "must have made an awful lot of numerical kinds of value, and "
                "then specified how they multiply so that one of them became "
                "weirdly tricky. Can you simplify?");
            break;
        case DimensionsInconsistent_KINDERROR:
            StandardProblems::sentence_problem(Task::syntax_tree(), _p_(PM_DimensionsInconsistent),
                "this is inconsistent with what is already known about those kinds of value",
                "all three of which already have well-established relationships - see the "
                "Kinds index for more.");
            break;
        case KindUnalterable_KINDERROR:
            Problems::quote_source(1, current_sentence);
            Problems::quote_source(2, pn);
            Problems::quote_kind(3, K1);
            Problems::quote_kind(4, K2);
            StandardProblems::handmade_problem(Task::syntax_tree(), _p_(PM_KindUnalterable));
            Problems::issue_problem_segment(
                "You wrote %1, but that seems to contradict %2, as %3 and %4 "
                "are incompatible. (If %3 were a kind of %4 or vice versa "
                "there'd be no problem, but they aren't.)");
            Problems::issue_problem_end();
            break;
        case KindsCircular_KINDERROR:
            Problems::quote_source(1, current_sentence);
            Problems::quote_source(2, pn);
            Problems::quote_kind(3, K1);
            Problems::quote_kind(4, K2);
            StandardProblems::handmade_problem(Task::syntax_tree(), _p_(PM_KindsCircular));
            Problems::issue_problem_segment(
                "You wrote %1, but that seems to contradict %2, as it would "
                "make a circularity with %3 and %4 each being kinds of the "
                "other.");
            Problems::issue_problem_end();
            break;
        case KindsCircular2_KINDERROR:
            Problems::quote_source(1, current_sentence);
            Problems::quote_source(2, pn);
            Problems::quote_kind(3, K1);
            Problems::quote_kind(4, K2);
            StandardProblems::handmade_problem(Task::syntax_tree(), _p_(PM_KindsCircular2));
            Problems::issue_problem_segment(
                "You wrote %1, but that seems to make %3 a kind of itself, which "
                "cannot make sense.");
            Problems::issue_problem_end();
            break;
        case LPCantScaleYet_KINDERROR:
            StandardProblems::sentence_problem(Task::syntax_tree(), _p_(PM_LPCantScaleYet),
                "this tries to scale up or down a value which so far has no point of "
                "reference to scale from",
                "which is impossible.");
            break;
        case LPCantScaleTwice_KINDERROR:
            StandardProblems::sentence_problem(Task::syntax_tree(), _p_(PM_LPCantScaleTwice),
                "this tries to specify the scaling for a kind of value whose "
                "scaling is already established",
                "which is impossible.");
            break;
        case NeptuneError_KINDERROR:
            Problems::quote_stream(1, E);
            StandardProblems::handmade_problem(Task::syntax_tree(), _p_(Untestable));
            Problems::issue_problem_segment(
                "One of the so-called Neptune files used to configure the kinds of value "
                "built into Inform contained an error. Either there is something wrong with "
                "this installation of Inform, or new Neptune files are being tried out but "
                "do not yet work. Specifically: '%1'.");
            Problems::issue_problem_end();
            break;
        default: internal_error("unimplemented problem message");
    }
}