OMGT Frontend Decoupling Assessment

1. Executive Summary

[!IMPORTANT] Strategic Reset (Jan 2026): Due to the high complexity identified below, the full migration of the core OrderMGT to Headless is POSTPONED INDEFINITELY. The strategy is now Hybrid: keep Kendo MVC for the core, use React/Headless only for new peripheral modules.

Decoupling the WepNG_OMGT frontend is a high-complexity and high-effort operation. The current architecture is a heavily coupled "Monolithic MVC Application" (.NET 4.8), where presentation and business logic are intertwined through Razor Views, ViewBag, Session, and legacy UI libraries (Kendo UI for MVC).

This is not a simple "restyling", but a complete re-architecture of the presentation layer.

2. Current State (Technical Analysis)

• Technology: ASP.NET MVC 5.2.7 on .NET Framework 4.8.
• Frontend Stack: jQuery 2.2.0, Bootstrap, Telerik Kendo UI for MVC.
• Coupling: CRITICAL.
  • Views (.cshtml) depend on server-side objects (html.Kendo().Grid(...)).
  • Data is passed via ViewBag and Model (Server-Side Rendering).
  • Navigation logic is managed by the server (RedirectToAction).
• Logic:
  • Controllers (e.g., OrderController.cs) are massive (> 1000 lines) and mix UI orchestration (redirects, TempData/ViewBag messages) with service calls.
  • JSON endpoints exist (e.g., GetLightData), which is positive, but they are a minority compared to HTML rendering.
• Dimensions:
  • The OrderMGT area alone counts ~50 Controllers and hundreds of views/partials.
  • Massive use of "Partial Views" makes clean component extraction difficult.

3. Recommended Strategy: "Strangler Fig Pattern"

I strongly advise against a "Big Bang Rewrite" (rewriting everything from scratch at once), as the risk of introducing regressions and blocking development for months/years is too high.

Gradual Approach (Hybrid):

1. Pilot (Vertical Slice): Choose a single isolated feature (e.g., "Light Search" or a specific order detail tab).
2. API Layer: Expose necessary data via WepNG_API (or new API controllers in OMGT returning only JSON).
3. Modern Frontend: Introduce a modern framework (React, Vue, or Angular) in a separate project (dedicated repo or folder).
4. Embedding: Load the new "mini-app" inside the existing MVC View (emptying it of Razor content and leaving only the container div).
5. Iteration: Repeat module by module until the old MVC becomes just an empty shell, then remove it.

4. Cost Estimates (T-Shirt Sizing)

Estimates are expressed in terms of effort for a Team of 2-3 Senior Developers.

Phase	Description	Estimate (Time/Resources)	Complexity
Setup & Pilot	Framework choice, build pipeline setup (Webpack/Vite), Auth (JWT?), Implementation of 1st "Simple" Page.	1 Month	Medium
Order MGT Core	Migration of OrderController (Overview, Details, main Tabs). Removal of Kendo UI Grid (logic rewrite in JS).	6-9 Months	Very High
Configuration Area	Migration of management tables and configurations.	3-4 Months	Medium
Full Project	Completion of all areas (InboundMGT, OutboundMGT, etc.)	12-18 Months+	High

Total Estimated (OrderMGT + Core): ~6-9 Man-Months (to have a working version of the operating core).

Main Cost Drivers (What raises the price):

1. Kendo UI Migration: All "Free" logic offered by Kendo (filtering, sorting, grouping, excel export) must be rewritten or licensed on a new JS library (e.g., AG Grid, TanStack Table).
2. Validation Logic: Validation (currently server-side DataAnnotations) must be replicated on the client.
3. Session Management: Moving from Server Session to Stateless (Token) requires backend refactoring.

5. Technical Prerequisites

Before starting, it is necessary to:

• Decide on the Backend Stack (React/Vue/Angular).
• Define the Authentication strategy (OAuth2/OpenID Connect) to support SPA + API.
• Create a Modern Build Environment (Node.js, CI/CD for the frontend).