diff --git a/crates/arco-cli/src/inspect.rs b/crates/arco-cli/src/inspect.rs
index 9d523a3..acae782 100644
--- a/crates/arco-cli/src/inspect.rs
+++ b/crates/arco-cli/src/inspect.rs
@@ -272,7 +272,6 @@ fn build_set_records(program: &SemanticProgram, filtered_set_names: &[&str]) ->
 }
 
 fn find_set_alias(program: &SemanticProgram, set_name: &str) -> Option<String> {
-    // Check variable families for index aliases that map to this set
     for family in &program.variable_families {
         for (index, domain) in &family.index_domains {
             if domain == set_name && index != set_name {
@@ -280,7 +279,6 @@ fn find_set_alias(program: &SemanticProgram, set_name: &str) -> Option<String> {
             }
         }
     }
-    // Check constraint generation bindings
     for constraint in &program.active_constraints {
         for binding in &constraint.generation_bindings {
             if binding.domain == set_name && binding.variable != set_name {
@@ -295,11 +293,9 @@ fn infer_set_dtype(resolved: &arco_kdl::semantic::ResolvedSet) -> String {
     if resolved.values.is_empty() {
         return "string".to_string();
     }
-    // Check if all values are integers
     if resolved.values.iter().all(|v| v.parse::<i64>().is_ok()) {
         return "int".to_string();
     }
-    // Check if all values are floats
     if resolved.values.iter().all(|v| v.parse::<f64>().is_ok()) {
         return "float64".to_string();
     }
@@ -893,7 +889,6 @@ fn collect_term_refs_from_expr(
                 })
                 .collect();
 
-            // Check if this target is already in the output
             if !out.iter().any(|r| r.name == *target) {
                 out.push(TermRef {
                     name: target.clone(),
@@ -1064,7 +1059,6 @@ fn build_objective_terms(
                 });
             }
             Expr::Reduction(reduction) => {
-                // Check if the body references named expressions
                 let body_terms = split_additive_terms(&reduction.body);
                 let mut found_expressions = false;
                 for body_term in &body_terms {
diff --git a/crates/arco-cli/tests/example_cli_commands.rs b/crates/arco-cli/tests/example_cli_commands.rs
index a3edeb6..75c2917 100644
--- a/crates/arco-cli/tests/example_cli_commands.rs
+++ b/crates/arco-cli/tests/example_cli_commands.rs
@@ -46,7 +46,6 @@ fn inspect_produces_valid_toml() {
     let stdout = String::from_utf8_lossy(&output.stdout);
     let parsed: toml::Value = toml::from_str(&stdout).expect("output should be valid TOML");
 
-    // Check top-level sections exist
     assert!(parsed.get("meta").is_some(), "should have meta section");
     assert!(parsed.get("set").is_some(), "should have set section");
     assert!(
@@ -84,7 +83,6 @@ fn inspect_json_produces_valid_json() {
 
     let payload: Value = serde_json::from_slice(&output.stdout).expect("valid inspect json");
 
-    // Check structure matches TOML layout
     assert!(payload.get("meta").is_some());
     assert!(payload.get("set").is_some());
     assert!(payload.get("variable").is_some());
@@ -92,7 +90,6 @@ fn inspect_json_produces_valid_json() {
     assert!(payload.get("constraint").is_some());
     assert!(payload.get("objective").is_some());
 
-    // Check counts
     let counts = payload["meta"]["counts"]
         .as_object()
         .expect("counts object");
diff --git a/crates/arco-kdl/src/semantic/error.rs b/crates/arco-kdl/src/semantic/error.rs
index 7984280..70c3df6 100644
--- a/crates/arco-kdl/src/semantic/error.rs
+++ b/crates/arco-kdl/src/semantic/error.rs
@@ -143,7 +143,7 @@ pub enum SemanticError {
     #[diagnostic(
         code(arco::semantic::unresolved_rule_set_filter_identifier),
         help(
-            "if token is a categorical value, quote it in filter, e.g. `where {{ a == \"north\" }}`"
+            "if token is a categorical value, quote it in filter, e.g. `filter {{ a == \"north\" }}`"
         )
     )]
     UnresolvedRuleSetFilterIdentifier {
diff --git a/crates/arco-kdl/src/source/parser.rs b/crates/arco-kdl/src/source/parser.rs
index 4fd95e2..7119daf 100644
--- a/crates/arco-kdl/src/source/parser.rs
+++ b/crates/arco-kdl/src/source/parser.rs
@@ -165,15 +165,15 @@ fn parse_data(node: &KdlNode, context: &ParseContext<'_>) -> Result<DataDecl, So
 
     Ok(DataDecl {
         name: first_arg_string(node, 0, context)?,
-        source: optional_property_string(node, "source", context)?
-            .or(optional_property_string(node, "from", context)?)
-            .ok_or_else(|| SourceError::MissingProperty {
+        source: optional_property_string(node, "source", context)?.ok_or_else(|| {
+            SourceError::MissingProperty {
                 node: node.name().value().to_string(),
                 property: "source",
                 path: context.path.to_path_buf(),
                 source_text: Box::new(context.source_text.clone()),
                 span: node.span(),
-            })?,
+            }
+        })?,
         maps,
         sets,
         indices,
@@ -210,6 +210,13 @@ fn parse_param(
         .transpose()
         .map_err(|error| algebra_error(node, error.to_string(), context))?;
 
+    for child in node.iter_children() {
+        match child.name().value() {
+            "index" | "filter" | "reduce" => {}
+            other => return Err(unsupported_declaration_error(child, other, context)),
+        }
+    }
+
     Ok(crate::source::ParamDecl {
         name: first_arg_string(node, 0, context)?,
         value: positional_value(node, &indices, context)?,
@@ -264,7 +271,7 @@ fn parse_set(node: &KdlNode, context: &ParseContext<'_>) -> Result<SetDecl, Sour
                     domain: Some(domain),
                 });
             }
-            "filter" | "where" => {
+            "filter" => {
                 filter_expression = Some(algebra_text_from_node(child, context)?);
             }
             member => {
@@ -562,31 +569,17 @@ data "D" source="file.csv" {
     }
 
     #[test]
-    fn data_from_property_parses_same_as_source() {
+    fn data_from_property_is_rejected() {
         let path = PathBuf::from("test.kdl");
 
-        let with_source = r#"
-data "D" source="file.csv" {
-  map "X" from="name"
-}
-"#;
-
         let with_from = r#"
 data "D" from="file.csv" {
   map "X" from="name"
 }
 "#;
 
-        let source_parsed = parse_program_text(with_source, &path).expect("source= syntax parses");
-        let from_parsed = parse_program_text(with_from, &path).expect("from= syntax parses");
-
-        assert_eq!(source_parsed.program.data.len(), 1);
-        assert_eq!(from_parsed.program.data.len(), 1);
-        assert_eq!(
-            source_parsed.program.data[0].source,
-            from_parsed.program.data[0].source
-        );
-        assert_eq!(from_parsed.program.data[0].source, "file.csv");
+        let error = parse_program_text(with_from, &path).expect_err("from= should be rejected");
+        assert!(error.to_string().contains("source"));
     }
 
     #[test]
@@ -610,30 +603,19 @@ data "D" from="file.csv" {
     }
 
     #[test]
-    fn where_keyword_parses_same_as_filter() {
+    fn where_keyword_is_rejected() {
         let path = PathBuf::from("test.kdl");
 
         let with_where = r#"set "thermal" { in "gen"; where { type == "thermal" } }"#;
-        let with_filter = r#"set "thermal" { in "gen"; filter { type == "thermal" } }"#;
 
-        let where_parsed = parse_program_text(with_where, &path).expect("where syntax parses");
-        let filter_parsed = parse_program_text(with_filter, &path).expect("filter syntax parses");
-
-        assert_eq!(where_parsed.program.sets.len(), 1);
-        assert_eq!(filter_parsed.program.sets.len(), 1);
-
-        assert_eq!(
-            where_parsed.program.sets[0].filter_expression,
-            filter_parsed.program.sets[0].filter_expression
-        );
-        assert!(where_parsed.program.sets[0].filter_expression.is_some());
+        parse_program_text(with_where, &path).expect_err("where syntax should fail");
     }
 
     #[test]
     fn mixed_old_and_new_syntax_parses() {
         let path = PathBuf::from("test.kdl");
 
-        // Mix of old (control, map, alias=, filter) and new (var, alias, as=, where)
+        // Mix of legacy-compatible and canonical forms we still support.
         let mixed = r#"
 data "D" source="file.csv" {
   map "old_col" from="x"
@@ -654,16 +636,13 @@ scenario "S" { use "M" }
 
         let parsed = parse_program_text(mixed, &path).expect("mixed syntax parses");
 
-        // Verify data block
         assert_eq!(parsed.program.data.len(), 1);
         assert_eq!(parsed.program.data[0].maps.len(), 2);
 
-        // Verify sets
         assert_eq!(parsed.program.sets.len(), 2);
         assert_eq!(parsed.program.sets[0].alias, Some("o".to_string()));
         assert_eq!(parsed.program.sets[1].alias, Some("n".to_string()));
 
-        // Verify model controls (both var and control)
         assert_eq!(parsed.program.models.len(), 1);
         assert_eq!(parsed.program.models[0].controls.len(), 2);
     }
diff --git a/crates/arco-kdl/src/source/parser_helpers.rs b/crates/arco-kdl/src/source/parser_helpers.rs
index e9ee21d..1e6da8b 100644
--- a/crates/arco-kdl/src/source/parser_helpers.rs
+++ b/crates/arco-kdl/src/source/parser_helpers.rs
@@ -51,7 +51,7 @@ pub(super) fn parse_optional_filter_expression(
     context: &ParseContext<'_>,
 ) -> Result<Option<String>, SourceError> {
     for child in node.iter_children() {
-        if matches!(child.name().value(), "filter" | "where") {
+        if child.name().value() == "filter" {
             return Ok(Some(algebra_text_from_node(child, context)?));
         }
     }
@@ -91,7 +91,7 @@ pub(super) fn declaration_indices(
 
     for child in node.iter_children() {
         match child.name().value() {
-            "filter" | "where" | "bounds" => {}
+            "filter" | "bounds" => {}
             "index" => {
                 let index_name = first_arg_string(child, 0, context)?;
                 let domain = child
diff --git a/crates/arco-kdl/src/source/surface.rs b/crates/arco-kdl/src/source/surface.rs
index 878d9db..c8604b0 100644
--- a/crates/arco-kdl/src/source/surface.rs
+++ b/crates/arco-kdl/src/source/surface.rs
@@ -26,7 +26,6 @@ fn rewrite_math_block_at(text: &str, start: usize) -> Option<(String, usize)> {
         b'm' => rewrite_math_block(text, start, "minimize")
             .or_else(|| rewrite_math_block(text, start, "maximize")),
         b'u' => rewrite_math_block(text, start, "upper"),
-        b'w' => rewrite_math_block(text, start, "where"),
         _ => None,
     }
 }
@@ -119,7 +118,7 @@ fn rewrite_math_block(text: &str, start: usize, keyword: &str) -> Option<(String
         "constraint" => format!("{header} expression={encoded_body}"),
         "expression" => format!("{header} {{ formula {encoded_body} }}"),
         "minimize" | "maximize" => format!("{header} expression={encoded_body}"),
-        "lower" | "upper" | "if" | "filter" | "where" => {
+        "lower" | "upper" | "if" | "filter" => {
             format!("{header} expression={encoded_body}")
         }
         _ => return None,
diff --git a/crates/arco-kdl/tests/compile_suite.rs b/crates/arco-kdl/tests/compile_suite.rs
index a70986f..9511da1 100644
--- a/crates/arco-kdl/tests/compile_suite.rs
+++ b/crates/arco-kdl/tests/compile_suite.rs
@@ -480,7 +480,7 @@ data "links" source="data/links.csv" {
     index "i" { in "tech" }
     index "g" { in "generators" }
     index "b" { in "buses" }
-    where { feasible > 0 }
+    filter { feasible > 0 }
   }
 }
 
@@ -548,7 +548,7 @@ data "links" source="data/links.csv" {
     index "i" { in "tech" }
     index "g" { in "generators" }
     index "b" { in "buses" }
-    where { feasible > 0 }
+    filter { feasible > 0 }
   }
 }
 
@@ -621,7 +621,7 @@ data "links" source="data/links.csv" {
     index "i" { in "tech" }
     index "g" { in "generators" }
     index "b" { in "buses" }
-    where { feasible > 0 }
+    filter { feasible > 0 }
   }
 }
 
@@ -630,7 +630,7 @@ set "feasible_links" {
   index "i" { in "tech" }
   index "g" { in "generators" }
   index "b" { in "buses" }
-  where { a == "1" }
+  filter { a == "1" }
 }
 
 model "TupleDispatch" {
@@ -698,7 +698,7 @@ data "links" source="data/links.csv" {
     index "i" { in "tech" }
     index "g" { in "generators" }
     index "b" { in "buses" }
-    where { feasible > 0 }
+    filter { feasible > 0 }
   }
 }
 
@@ -706,7 +706,7 @@ set "target_pairs" {
   in "feasible_links"
   index "a" { in "area" }
   index "i" { in "tech" }
-  where { generators == "g1" }
+  filter { generators == "g1" }
 }
 
 model "TupleDispatch" {
@@ -782,7 +782,7 @@ data "links" source="data/links.csv" {
     index "i" { in "tech" }
     index "g" { in "generators" }
     index "b" { in "buses" }
-    where { feasible > 0 }
+    filter { feasible > 0 }
   }
 }
 
@@ -850,7 +850,7 @@ data "links" source="data/links.csv" {
     index "i" { in "tech" }
     index "g" { in "generators" }
     index "b" { in "buses" }
-    where { feasible > 0 }
+    filter { feasible > 0 }
   }
 }
 
@@ -916,7 +916,7 @@ data "links" source="data/links.csv" {
     index "i" { in "tech" }
     index "g" { in "generators" }
     index "b" { in "buses" }
-    where { feasible > 0 }
+    filter { feasible > 0 }
   }
 }
 
@@ -1073,7 +1073,7 @@ data "links" source="data/links.csv" {
     index "i" { in "tech" }
     index "g" { in "generators" }
     index "b" { in "buses" }
-    where { feasible > 0 }
+    filter { feasible > 0 }
   }
 }
 
diff --git a/crates/arco-kdl/tests/semantic_validation.rs b/crates/arco-kdl/tests/semantic_validation.rs
index d34c8e8..61a36d9 100644
--- a/crates/arco-kdl/tests/semantic_validation.rs
+++ b/crates/arco-kdl/tests/semantic_validation.rs
@@ -649,7 +649,7 @@ data "links" source="data/links.csv" {
     index "i" { in "tech" }
     index "g" { in "generators" }
     index "b" { in "buses" }
-    where { feasible > 0 }
+    filter { feasible > 0 }
   }
 }
 
@@ -692,7 +692,7 @@ set "tech" { "wind"; "solar" }
 set "feasible_links" {
   index "a" { in "area" }
   index "i" { in "tech" }
-  where { unknown_col == "1" }
+  filter { unknown_col == "1" }
 }
 
 model "Dispatch" {
@@ -894,7 +894,7 @@ data "links" source="data/links.csv" {
     index "i" { in "tech" }
     index "g" { in "generators" }
     index "b" { in "buses" }
-    where { feasible > 0 }
+    filter { feasible > 0 }
   }
 }
 
diff --git a/docs/appendix-a-ergonomic-profile.md b/docs/appendix-a-ergonomic-profile.md
index 93e0b5b..f5c06a6 100644
--- a/docs/appendix-a-ergonomic-profile.md
+++ b/docs/appendix-a-ergonomic-profile.md
@@ -51,7 +51,7 @@ canonical `set { in ...; filter { ... } }` syntax defined in
 This is a usage guide, not new grammar.
 
 ```kdl
-data branch_data from="data/branches.csv" {
+data branch_data source="data/branches.csv" {
   set edge
   set active_edge { in edge; filter { conex == 1 } }
   param conex index=edge
@@ -82,7 +82,7 @@ data branch_data from="data/branches.csv" {
 >     index i { in tech }
 >     index g { in generators }
 >     index b { in buses }
->     where { feasible > 0 }
+>     filter { feasible > 0 }
 >   }
 > }
 >
@@ -92,7 +92,7 @@ data branch_data from="data/branches.csv" {
 >   index i { in tech }
 >   index g { in generators }
 >   index b { in buses }
->   where { area == "south" }
+>   filter { area == "south" }
 > }
 >
 > model nodal_allocation {
diff --git a/docs/arco-spec.md b/docs/arco-spec.md
index 4ec2c5c..3c6b358 100644
--- a/docs/arco-spec.md
+++ b/docs/arco-spec.md
@@ -392,17 +392,16 @@ Required properties:
 
 - `source`: CSV file path. Relative paths are resolved from the directory
   containing the `.kdl` file being parsed. Absolute paths are also accepted.
-  - `from`: compatibility alias for `source` (legacy syntax). Prefer `source` in
-    new files.
-    CSV parsing: Arco expects RFC 4180-compliant CSV files (comma-delimited,
-    optional double-quote escaping, CRLF or LF line endings). The first row MUST be
-    a header row containing column names. Column names in the header MUST be unique;
-    duplicate column names MUST fail validation (see
-    [§10](#10-validation-requirements), rule 73). CSV files MUST be UTF-8 encoded.
-    Implementations SHOULD accept files with or without a UTF-8 BOM. Empty cells in
-    a numeric column MUST fail validation. Empty cells in a string/categorical
-    column are treated as empty strings. Column matching is always by name, not by
-    position.
+
+CSV parsing: Arco expects RFC 4180-compliant CSV files (comma-delimited,
+optional double-quote escaping, CRLF or LF line endings). The first row MUST be
+a header row containing column names. Column names in the header MUST be unique;
+duplicate column names MUST fail validation (see
+[§10](#10-validation-requirements), rule 73). CSV files MUST be UTF-8 encoded.
+Implementations SHOULD accept files with or without a UTF-8 BOM. Empty cells in
+a numeric column MUST fail validation. Empty cells in a string/categorical
+column are treated as empty strings. Column matching is always by name, not by
+position.
 
 Allowed children:
 
diff --git a/examples/nodal-allocation/input.kdl b/examples/nodal-allocation/input.kdl
index e90a383..3183a42 100644
--- a/examples/nodal-allocation/input.kdl
+++ b/examples/nodal-allocation/input.kdl
@@ -12,7 +12,7 @@ data "links" source="data/links.csv" {
     index "i" { in "tech" }
     index "g" { in "generators" }
     index "b" { in "buses" }
-    where { feasible > 0 }
+    filter { feasible > 0 }
   }
 
   param "capacity_mw" {
@@ -36,7 +36,7 @@ set "priority_links" {
   index "i" { in "tech" }
   index "g" { in "generators" }
   index "b" { in "buses" }
-  where { area == "south" }
+  filter { area == "south" }
 }
 
 model "NodalAllocation" {
diff --git a/scripts/.gitkeep b/scripts/.gitkeep
deleted file mode 100644
index e69de29..0000000
diff --git a/tools/tree-sitter-arco-kdl/examples/highlight_demo.kdl b/tools/tree-sitter-arco-kdl/examples/highlight_demo.kdl
index 93ba2d9..2790b44 100644
--- a/tools/tree-sitter-arco-kdl/examples/highlight_demo.kdl
+++ b/tools/tree-sitter-arco-kdl/examples/highlight_demo.kdl
@@ -6,7 +6,7 @@ set time alias=t {
   member 3
 }
 
-data generators from="data/generators.csv" {
+data generators source="data/generators.csv" {
   map gen from=generator
   param capacity_mw index=gen
   param fuel_cost from=cost_per_mwh index=gen