HANDALPUM public Site — REA reference analysis

Date: 2026-10-15 (Asia/Seoul). Analyst: REA evidence worker.

Scope and method

The six user-selected pages were inspected with REA 6.5.0, using capture_browser_scenario and a temporary, owned Chrome process. The first W capture was reused from the previous session. Five page captures and three focused click scenarios ran through the newly available REA MCP connection. All nine scenarios completed, and all owned browser processes reported terminated-owned-process cleanup.

Environment: Chromium 155.0.8059.79, 1440 × 900, ko-KR, Asia/Seoul. Each initial page capture waited 1,500 ms and retained URL, DOM, accessibility text and a viewport screenshot. Follow-ups declared their click and another 1,500 ms wait. No form was filled or submitted, no booking/payment was created, and no login was attempted. Screenshots from the six initial captures were visually inspected.

Coverage: six supplied pages, plus W's VIP tab and Goongs' reservation landing page. This is sufficient reference evidence for the public demo's information architecture and visual direction, not a complete reconstruction of six production sites. Capture completeness describes the requested artifacts; it does not imply every site route, animation or browser state was inspected.

Evidence index

The adjacent compact manifest preserves capture identities, artifact digests/byte sizes, request parameters, outcomes and local raw-file hashes. Full payloads remain in /tmp/postpartum-rea-evidence/: w-woman.json (E1) and site-evidence-bundle.json (E2–E9). The latter contains eight records, 13,586,832 bytes. These temporary local payloads are not published with the demo and are not durable repository artifacts. No third-party screenshots, site source, contact details, prices or medical claims are shipped as HANDALPUM content.

Direct observations

Each statement below describes captured content or pixels. Published claims about care or staff are not evidence of clinical quality.

W — room comparison and the stay schedule (E1, E7)

• The page sequence is introduction, care differentiators, facilities, program schedule, prices and contacts. The initial viewport uses a room photograph under a hospital navigation bar, blue accents and a right-side contact rail.
• Four facility buttons separate three room categories and other facilities. Each panel contains an iframe. A schedule and room-by-duration price table provide detailed comparison content.
• Runtime verified: after the declared VIP click, the captured DOM has the VIP button's active class, #tab3 displayed and #tab1, #tab2, #tab4 hidden. The parent URL remains /66. This upgrades the earlier static customTabOpen observation to an observed state transition. Inner iframe content was not separately captured.

Seoul Women's Hospital — practical decision details (E2)

• Hospital-wide navigation and a left section menu surround the use guide. The screenshot shows magenta accents, a narrow reading column and a room banner; reservation information starts within the initial viewport.
• The page explains reservation timing, arrival/departure, room allocation, room/duration pricing, included services, add-ons and preparation items. A substantial terms section follows. A dedicated enquiry-board link exists.
• The displayed tariff includes an effective date. Neither current availability nor the actual enquiry-board submission process was tested.

Versailles — hospitality presentation and enquiry (E3)

• The initial viewport has an oversized visual hero, overlaid navigation, language controls and large white text. The captured background shows a city video frame with visible third-party player/title overlays.
• Below the hero, the accessible document describes introduction, programs, facility categories, usage information and an enquiry form. Facility groups and six slide buttons are present. A language selector offers nine named languages; translation behavior was not exercised.
• The form exposes name, email, contact number, expected date and message fields with a send button. Delivery, validation, consent, persistence and back-office processing remain untested.

Goongs Gangnam — visual identity and reservation entry (E4, E8)

• A full-height exterior photograph, centered identity and top contact text dominate the initial screenshot. The accessible document also has quick routes to reservation, Q&A and charges, followed by distinct care themes.
• Runtime verified: clicking the reservation quick link changes the URL to /reservation/reserv. The destination exposes consultation instructions, duration, departure guidance, room/program tariffs, included services and add-ons. No form submission is part of this transition.
• Facility-led brand presentation and easily identifiable practical information coexist on separate parts of the site. Conversion performance is unknown.

Marie The Bless — care categories and persistent contact (E5, E9)

• The initial screenshot has a room/lobby photograph, centered typography, understated navigation and a right-side rail for charges, social media, online enquiry and telephone contact. Carousel controls are visible.
• The document separates facilities, newborn care, mother care, spa, reviews and directions. The review section contains an access-required message in this unauthenticated capture; it does not establish absence of reviews.
• The online-enquiry rail link declares target="_blank". The click completed, but E9's captured original page stayed at /. Popup events/content were not selected, so the destination and submission flow are unknown. The click is not evidence that an enquiry form loaded or that the link is broken.

Yoonho — guide hierarchy and transparent inclusions (E6)

• The screenshot uses a broad soft photographic banner, restrained white content area, clear section labels and guide subnavigation.
• The captured guide separates stay duration, reservation approach, room categories, basic services, extra services, room-use conditions and changes or cancellation. Arrival/departure and visiting guidance have separate links.
• Consultation is described as a telephone-led process. Neither live terms, fees nor room availability were verified; no policy text is adopted as legal guidance for the fictional demo.

Inferences and implementation decisions

These are design judgments derived from the observations, not measured visitor behavior or claims of conversion uplift.

Observation-to-CoUI boundary

CoUI provides the shared theme, button, text field and disclosure behavior. The composition is a Jaspr public page with responsive semantic HTML; native select/checkbox controls preserve normal static-page form semantics. Large editorial photography and content spacing are application composition decisions. A toolkit-wide limitation or need for a new CoUI primitive cannot be established from competitor screenshots alone; any library change needs an observed issue in the running HANDALPUM page and a focused upstream reproduction.

The public experience is intentionally a fictional enquiry simulation. No source site's form endpoint, medical claims, logo, copyrighted room image, price, bank account or real contact destination is reused. The existing CRM Console is linked as a separate demo surface; no successful live reservation is implied by a locally rendered receipt.

Remaining unknowns and verification boundary

1. Only the selected desktop viewport was observed on the source sites. Source mobile behavior, screen-reader completeness, performance and network/API mechanisms remain unmeasured.
2. No login, real enquiry, payment, submission or medical service was exercised.
3. W's embedded room contents and The Bless's new-tab destination were not inspected. Third-party frame and service behavior is outside this capture.
4. Original carousel timing, all tabs, language switching and form validation are not exhaustively tested. DOM presence is distinct from execution.
5. REA event ordering does not prove causality. Selected network events/bodies were not retained; there is no API or request-origin conclusion here.
6. HANDALPUM build, accessibility, 11-locale and visual verification belongs to the implementation test report. This reference-analysis report does not assert that the new site has passed those checks or been published.

Checklist: six page identities/artifacts answered; information architecture and desktop visual patterns answered within capture scope; representative room/reservation transitions partially answered (E7/E8); source live booking workflows unexamined by design; new-site implementation checks separate.