README.md

@@ -312,9 +312,9 @@ order.id: datatype integer: "1{digits(2)}" is not one value; write one literal o
 
 A column of one reference alone to another record's column — `"score": "{/src.score}"`
 — is that column: it takes the column's datatype and is null where the column is, and
-a struct field tagged `src.score` is nil there. A `datatype` of its own types a string
+a struct field tagged `src.score` is nil there. A `datatype` of its own types the
-column's values, and over a typed column is refused naming the datatype it takes. Any
+column's values where they prove it, and one restating the datatype it takes is refused.
-other read renders the column's text, a null as `""`.
+Any other read renders the column's text, a null as `""`.
 
 A typed column's `{calc()}` must be proven to print a number: each operand a number
 literal, an `{int()}`, `{float()}`, `{seq()}` or `{digits()}` call, a calc, or a read of
@@ -549,10 +549,11 @@ tokens add cost in proportion to the output.
   tag of one reference alone, `{/users}`, is refused naming the path `users`: both
   render the same text, and only the path names a record. `IsTemplate` exports the
   rule, so the CLI, struct tags and any other caller read one.
-- **An inline template skips the cycle fence, and only that one.** `New` proves the
+- **An inline template skips the cycle fence.** `New` proves the loaded tree
-  loaded tree acyclic, an inline node is a finite tree of its own, and nothing in
+  acyclic, an inline node is a finite tree of its own, and nothing in the tree can
-  the tree can reference it, so no render of it reaches itself. Every other fence
+  reference it, so no render of it reaches itself. Every other fence runs over both,
-  runs over both, from one `checkScope`.
+  from one `checkScope`, except that struct tags leave column agreement to their Go
+  types.
 - **An inline template that does not compile is misuse (exit 2), including a
   reference that resolves to nothing** — the whole argument is the spelling under
   test, and `NewTemplate` compiles, links and validates as one step. An unknown
@@ -672,9 +673,9 @@ tokens add cost in proportion to the output.
   one comparison, and `1{digits(2)}` is a second spelling of `{int(100,199)}`.
 - **A column of one reference alone is the column it reads.** `{/src.score}` renders
   exactly what `src.score` draws, so it takes that column's datatype and null rather
-  than restating them, and a `datatype` restating a typed column is a second spelling.
+  than restating them, and a `datatype` restating the one it takes is a second
-  Over a string column a `datatype` still types the values — the one way to type a
+  spelling. Any other `datatype` still types the values — the one way to type a column
-  column someone else wrote.
+  someone else wrote.
 - **A typed column's calc is refused unless proven.** Operand bounds must keep each
   divisor from zero and the result finite; what they cannot show is refused rather
   than trusted, since a bare `NaN` breaks the JSON and SQL it lands in.

datatype.go

@@ -165,11 +165,33 @@ func disagreement(a *template, da DataType, b *template, db DataType) error {
 	}
 	switch {
 	case da != DataTypeString && db != DataTypeString:
-		return fmt.Errorf("its items hold %s and %s; a column holds one datatype", da, db)
+		return bothTyped(a, da, b, db)
 	case typed.datatype == DataTypeString && (&valueProof{}).columnItem(bare).not[want] != "":
-		return fmt.Errorf(`item %q is not %s, the datatype item %q takes from the column it reads; to read that column as text, write {"format":"{text}","text":%q}`, bare.format, dataTypeNouns[want], typed.format, typed.format)
+		return fmt.Errorf(`item %q is not %s, the datatype item %q takes from the column it reads; to read that column as text, %s`, bare.format, dataTypeNouns[want], typed.format, asText(typed))
 	case bare.fromString: // an object may carry a weight, which this spelling would drop
 		return fmt.Errorf(`item %q declares no datatype, and a column holds one; write it as {"format":%q,"datatype":%q}`, bare.format, bare.format, want)
 	}
 	return fmt.Errorf(`item %q declares no datatype beside one declaring %s; a column holds one, so give it "datatype": %q`, bare.format, want, want)
 }
+
+// bothTyped names the fix for two items holding different datatypes: reading one that takes its
+// datatype from the column it reads as text, since only a declared datatype can be edited away.
+func bothTyped(a *template, da DataType, b *template, db DataType) error {
+	read := a
+	if a.datatype != DataTypeString {
+		read = b
+	}
+	if read.datatype != DataTypeString {
+		return fmt.Errorf("its items hold %s and %s; a column holds one datatype", da, db)
+	}
+	return fmt.Errorf("its items hold %s and %s; a column holds one datatype, so to read %q as text, %s", da, db, read.format, asText(read))
+}
+
+// asText names the spelling that reads the column a column-read item reads as text, keeping the
+// other keys an object item carries.
+func asText(t *template) string {
+	if t.fromString {
+		return fmt.Sprintf(`write {"format":"{text}","text":%q}`, t.format)
+	}
+	return fmt.Sprintf(`set its "format" to "{text}" and add "text": %q`, t.format)
+}

graph.go

@@ -177,7 +177,6 @@ func inlineScope(n node, label string) nodeScope {
 	return func(fn func(path string, m node) error) error { return eachNode(n, label, fn) }
 }
 
-// checkScope runs every fence over a scope: checkColumns, then checkRenders.
 func checkScope(s nodeScope) error {
 	if err := checkColumns(s); err != nil {
 		return err

value.go

@@ -26,13 +26,13 @@ type valueProof struct {
 }
 
 // checkDatatype rejects a typed column item some render of which is not text of its datatype,
-// and one declaring a datatype over the typed column it is.
+// and one restating the datatype of the column it is.
 func (p *valueProof) checkDatatype(path string, n node) error {
 	t, ok := n.(*template)
 	if !ok || t.datatype == DataTypeString {
 		return nil
 	}
-	if d := readDatatype(t); d != DataTypeString {
+	if d := readDatatype(t); d == t.datatype {
 		return fmt.Errorf(`%s: %s takes datatype %s from the column it reads; drop "datatype"`, path, t.format, d)
 	}
 	if reason := p.columnItem(t).not[t.datatype]; reason != "" {