Acknowledge 'sentence unclear' feedback instead of closing silently

CHANGELOG.md

@@ -3,6 +3,7 @@
 This document summarizes notable updates since February 2025, with commit dates from the repository history for context.【728428†L1-L48】
 
 ## 2026-07
+- **UX (2.1.3):** Submitting the "Is this English sentence confusing or unclear?" report on an english-to-LTL question now acknowledges the feedback instead of silently closing the modal: a persistent note appears under the question ("Thank you for your feedback. Please answer with your best-guess interpretation of the sentence."), announced to screen readers via `role="status"`. A failed submission now alerts the student instead of logging only to the console.
 - **UX (2.1.2):** Misconception explainers now show inline LTL operators and formulas in the learner's selected Classic, Forge, or Electrum syntax, updating immediately when the top-bar syntax choice changes. Explainer fragments are authored in Classic inside `<code class="ltlformula">` and rewritten client-side by token substitution — safe because the tutor grammar only permits lowercase atoms, so a bare `X`/`F`/`G`/`U` is always an operator.
 - **Bugfix (2.1.2):** The precedence explainer claimed `a | b & c -> d` parses as `((a | b) & c) -> d`; the parser actually yields `(a | (b & c)) -> d` (binary operators bind `&` > `|` > `U` > `->` > `<->`, left-associative within a level). Separately, the Electrum rendering of `U` formulas rendered operands with the Forge renderer, so e.g. `(X p) U q` displayed `NEXT_STATE` where Electrum uses `AFTER`.
 - **Experiment (2.1.1):** The english-to-LTL framing experiment now has three arms (one third each) instead of the 50/50 abstract/lights split: abstract, lights (concrete but *descriptive*), and a new deontic arm that frames the *same* formula as a document-access policy the student audits ("You are auditing access to a confidential document. Company policy: Whenever the document is open, the VPN must be connected."). Per the Wason selection-task literature, facilitation comes from deontic, violation-checkable rules rather than concrete content alone — so the lights arm becomes the mere-content control. The audit theme's four attributes (document open, clearance active, VPN connected, screen shared) are deliberately independent in a student's mental model: coupled only by the policy under test, never by physics, so every state assignment stays conceivable. Deontic phrasing places the obligation on consequents only ("is connected" → "must be connected"); triggers stay indicative. New responses log `translation_mode="contextualizeddeontic"`; the lights arm keeps its historical `"contextualized"` value and its exact wording, so earlier responses stay comparable.

src/templates/exercise.html

@@ -86,6 +86,9 @@ <h5 id="{{ qindex }}question_text" class="card-title mb-2">
                             <button type="button" class="btn btn-link text-primary p-0" data-toggle="modal" data-target="#ltlToEngRaterModal-{{ qindex }}">
                                 Is this English sentence confusing or unclear?
                             </button>
+                            <p id="unclear-ack-{{ qindex }}" class="w-100 text-muted mb-0 mt-1 d-none" role="status">
+                                <small>Thank you for your feedback. Please answer with your best-guess interpretation of the sentence.</small>
+                            </p>
                             {% include 'ltltoengrater.html' %}
                         {% endif %}
                     </div>

src/templates/ltltoengrater.html

@@ -127,10 +127,13 @@ <h5 class="modal-title" id="ltlToEngRaterModalLabel-{{ qindex }}">What issues di
                 });
 
                 if (response.ok) {
-                    // Hide the modal on success
                     $(`#ltlToEngRaterModal-${qindex}`).modal('hide');
+                    const ack = document.getElementById(`unclear-ack-${qindex}`);
+                    if (ack) {
+                        ack.classList.remove('d-none');
+                    }
                 } else {
-                    console.log('Failed to submit feedback. Please try again.');
+                    alert('Failed to submit feedback. Please try again.');
                 }
             } catch (error) {
                 console.error('Error submitting feedback:', error);

src/templates/version.html

@@ -1 +1 @@
-2.1.2
+2.1.3