Assignment 4 | Prototype Design | Software_clan
Assignment 4 CS2 Smart Replay Platform
- Software Engineering Team Project
Format Description
| Item | Content |
|---|---|
| Course | Software Engineering (2601_FZU-MU_SE) |
| Assignment Requirements | https://bbs.csdn.net/topics/620556488 |
| Team Name | Software_clan |
| Objective of this assignment | Complete interface prototype design, create high-fidelity interactive prototype, write design report PPT, and conduct live demo. |
| Other Reference Documents | The Method of Construction: Modern Software Engineering (3rd Edition), Chapter 8 Requirements Analysis and Chapter 10 Typical Users and Scenarios |
Content Index
- 1. Team division of labor and contribution
- 2. Interface prototype design
- 3. Challenges and solutions
- 4. Takeaways and team evaluation
- 5. Team discussion photos (non-posed)
- 6. PPT link
1. Team division of labor and contribution
The prototype design phase (4th assignment) marked an important step for the team moving from requirements analysis toward product visualization. Unlike the 3rd assignment, which focused on document writing, the core deliverable of this assignment is an online-accessible, clickable high-fidelity interactive prototype, along with a defense PPT and blog documentation. Therefore, the division of labor shifted from "document chapter ownership" to "ownership by prototype module, presentation task, and blog section," ensuring every deliverable has a clear owner while avoiding overlap.
| Name & Student ID | Work Description | Contribution |
|---|---|---|
| Shixun Guo (832402102) | Organized presentation logic, created defense PPT, integrated prototype screenshots and demo flow. | 14% |
| Yiyao Guo (832402104) | Completed core prototype design: Responsible for Modao high-fidelity prototype construction, page interaction linking, hotzone configuration, and overall visual implementation. | 24% |
| Keyou Su (832402117) | Collected internal trial feedback, organized issue list, and served as on-site speaker. | 14% |
| Yan Zheng (832401326) | Responsible for integrating all sections and writing the final blog | 14% |
| Siqi Mo (832401314) | Responsible for writing the "Challenges & Solutions" section | 12% |
| Yuefan Xu (832401122) | Responsible for writing the "Prototype Design" section of the blog | 12% |
| Zhanhong Liu (832401312) | Responsible for teammate evaluation | 5% |
| Yanzhengyu Guo(832402103) | Responsible for teammate evaluation | 5% |
2. Interface prototype design
2.1 Prototype tool
Tool: Modao
For this prototype design, Modao was selected as the primary tool. During the tool selection phase, the team compared common prototyping tools such as Axure RP, Balsamiq Mockup, and Figma, and ultimately chose Modao based on the following four considerations:
(1) Online sharing, no installation
(2) Hotzone interaction, cross-page jumps
(3) Component reuse, visual consistency
(4) Team collaboration space
2.2 Design process
The prototype design was not accomplished overnight but went through six stages of progressive convergence. Each stage had clear deliverables, and the output of the previous stage served as the input for the next, avoiding the chaos of "drawing pages while changing requirements."
Stage 1: Requirements Analysis
Based on SRS V1.1, the team confirmed 22 functional requirements and 3 core use cases. The key at this stage was clarifying the prototype scope—which functions must be reflected in the prototype and which can be replaced by text annotations. For example, real Demo upload and backend parsing are not within the prototype scope, but the four-step parsing status on the "New Review" page must be displayed upfront in the prototype so reviewers understand the asynchronous flow of the real product.
Stage 2: Information Architecture
Determined 6 left navigation items (Overview, My Reviews, Aim Training, Campus Squad, Players & Friends, Personal Center) and a list of 10 pages. This stage solved the question of "whether the four modules can fit into the same workbench"—unified left navigation, login state, and page skeleton, avoiding each module being built separately.
Stage 3: Low-Fidelity
Drew wireframe sketches based on Chapter 7 of the SRS, focusing on confirming page layout and information hierarchy. The low-fidelity stage did not fuss over colors and fonts, only answering "what should be seen at first glance on this page."
Stage 4: High-Fidelity
Determined the dark theme, navigation, and component library. The UI teammate defined dark backgrounds, panel hierarchies, orange accent colors, button styles, and metric card styles, and made them reusable components. The key at the high-fidelity stage was visual consistency—10 pages placed together should not feel like they were made by different people.
Stage 5: Hotzone Interaction
Cross-page jumps + state annotations. This stage linked buttons, navigation, cards with real usage paths: Login → Overview → New Review → Report → My Reviews, as well as branches to training, squad, and friends from the overview. After hotzone configuration, the prototype changed from "static screenshots" to a "clickable demo" interactive prototype.
Stage 6: Review & Iteration
Prototype review, walkthrough, and revision. QA teammates walked through hotzones and field consistency against SRS use cases, UI and frontend teammates provided improvement suggestions from their perspectives, and revisions after review were synchronized to the blog. This stage ensured the prototype is not only "clickable" but "clickable correctly."
2.3 Screens and interactions
The prototype includes 10 high-fidelity pages covering four modules: Smart Replay, Campus Squad, Aim Training, and Account Center. Each page has a clear functional positioning and interaction path, described below by module.
Page 01 Overview
The Overview is the workbench homepage, responsible for "seeing status and next steps at a glance." The top of the page is the value proposition "Turn the last match into the next advantage," with the primary action button "+New Review" leading directly to the creation flow. The middle shows status: 3 matches reviewed, average Rating 1.18, average ADR 89.3, match win rate 67%. The bottom has three "Next Step" cards: analyze the last match, practice a set first, play together next match, leading to review, training, and squad respectively.

Page 02 - Login & Register
The Login & Register page is an entry point connecting the workbench and backend identity. The left side has the welcome message "Next progress starts here," and the right side has the account input area with username demo_player and password; the formal product verifies identity via backend, as annotated on the prototype page. Two actions: "Login and Enter" goes to the workbench, "Create Account" for new user registration. Permission boundary: review, squad, and personal center require login; aim training supports no-login access, with scores stored locally first.

Page 03 - My Reviews
My Reviews is the match archive page, with the core design principle "original Demo stays local." The top of the page prompts "Original Demo stays local; here shows the analysis report and history for each match," clarifying the boundary of "file locally, report on platform" to avoid players misunderstanding the upload behavior. The match list includes three recent matches: Mirage, Inferno, Ancient, with dates, scores, rounds, and status. Each record has "View Report" to enter the review, and the top allows creating a new review.

Page 04 - New Review
The New Review page puts the four-step flow upfront, with zero-threshold sample data. The page offers two import methods: automatic import and manual ZIP upload, with a drag-and-drop area to initiate analysis. Four-step flow: 01 Read data → 02 Identify rounds and players → 03 Calculate performance metrics → 04 Generate report. The page clearly annotates "The defense prototype uses sample data and does not upload real Demo." Parsing is an asynchronous process, and the prototype uses four-step states to display the real states of "queuing, parsing, completed, failed" upfront.

Page 05 - Review Report
The Review Report is the core page of FR-1 Smart Replay, and the report must strictly match the current match and player. Match header: Mirage / DE_MIRAGE, 2026/10/24, 23 rounds, our team 10:13 opponent. Key metrics: KDA 33/14/2, ADR 141.6, contribution round rate 78.3%, headshot rate 30.3%. Tabs and charts: four tabs of Match Overview, Round Review, Utility & Survival, Detailed Data; 23-round influence bar chart.

Page 06 - Aim Training
The Aim Training page shows five modes: Grid Tap, Reaction Test, Large Flick, Precision Micro-adjustment, Smooth Tracking, corresponding to FR-3 requirements. The prototype shows mode selection and operation entry, with a unified "Select Mode" action. The page annotates: the actual frontend training module contains the complete game; the prototype does not redraw the game screen. Aim training is a pure frontend module, retaining the original game running method; the prototype only does the entry and flow.

Page 07 - Campus Squad
The Campus Squad page turns team formation from QQ group word-of-mouth into a lobby list with filters, rank, and region information. Three-step flow: Get team code → Join in game → Captain starts, closing the campus scenario loop. Filters: All Modes, Competitive Match, Priority Mode, Partner Mode; 6 practice squad cards. Squad cards include map, rank, region; "+Publish Squad" initiates recruitment.

Page 08 - Players & Friends
The Players & Friends page supports searching players and finding teammates by specialty. The top search box supports finding players by nickname. The friend list includes Alice, Bob, Charlie, annotated with specialties in entry fragging, utility communication, and online time. Each friend has "View Card" to enter the player card page to understand level and needs. Friend cards make "role complementarity" explicit, providing a basis for squad matching.

Page 09 - Personal Center
The Personal Center page is the landing point of FR-4, linking player profile, friends, and squad together. Personal card: Demo Player, East China · Perfect C+, tags: Looking for Teammates, Mirage, Competitive Match. Profile editing: nickname, Steam friend code, personal signature, synchronized after saving. Two entries: View Friends goes to Players & Friends, Find Squad goes to Campus Squad.

Page 10 - Player Card
The Player Card page is a squad decision page, allowing players to quickly judge suitability before joining a squad. Player profile: Alice, East China · Perfect C+, tags Mirage, Looking for Teammates, Weekend Online. Game level Perfect C+, regular maps Mirage/Inferno. Teammate needs: friendly communication, stable play; can add friend or return to friend list. The card is a squad decision page, allowing players to quickly judge suitability before joining a squad.

Core flow:

2.4 Prototype Link
https://modao.cc/proto/c6pYF5D8tngg2yPkh90v/sharing?view_mode=device&screen=rbpVZ2KbgPN2vHiVI #CS Demo Master 界面原型-proto
3. Challenges and solutions
Core Challenges Encountered by the Team
The most challenging part of prototype design is integrating modules from different teammates into the same platform. For example, account and team functions need to interface with backend APIs; aim training must retain the original game running method; Demo review involves local parsing, upload, and report display—data sources and states are all different. The hardest part is verifying that the report matches the current match and player, while loading large map data in batches to avoid displaying wrong data or causing page lag.
(1) Difficulty 1: Cross-Module Integration—Different Data Sources and States
Problem Description: Account and squad functions need to interface with backend APIs; aim training retains the original game running method; review involves local parsing, upload, and report display.
Attempts & Solutions: Unified the workbench framework and left navigation; mapped module states (not logged in/parsing/analyzed/recruiting); annotated API fields and data flow with sample data in the prototype, aligning frontend and backend on the prototype.
(2) Difficulty 2: Data Consistency & Map Data Performance
Problem Description: The report must match the current match and player; the map data volume is large, and loading it all at once can easily display wrong data or cause lag.
Attempts & Solutions: Strongly bound the match header (map, date, score, player) to report data; lazy-loaded map data by round and player, rendered in batches, producing the report framework first before filling in charts.
4. Takeaways and team evaluation
4.1 Three Main Takeaways
(1) Determined the unified workbench and navigation first, then built individual pages; the four modules share the same skeleton, so subsequent integration will not require rework.
(2) Used sample data to first validate page states, button paths, and chart layouts, providing a clear reference standard for real feature development.
(3) Account, squad, and review functions annotated data sources and fields at the prototype stage, significantly reducing frontend-backend communication costs.
4.2 Real Feedback from Teammates
Zhanhong Liu · UI Perspective
- Dark design has an esports style; inverted display is highly readable and adapts to low-light environments
- Left menu bar is simple and clear
Keyou Su · Player Perspective
- Content is concise and easy to understand; even zero-basis players can understand what each part does
Yanzhengyu Guo · Frontend Perspective
- Black-orange color scheme fits the positioning; unified style with strong atmosphere
- Modular information; player basic info and introduction in two separate cards with clear hierarchy
- Orange "Add Friend" button is distinct and intuitive
4.3 Feedback Issues & Developer Responses
(1) Menu font too small, folds when window is small
Reported by: Zhanhong Liu
Resolution: Adjust menu font size and responsive breakpoints; switch to icon navigation + drawer expansion at small window widths to avoid incomplete display.
(2) Only dark theme, too dark; hope for light option
Reported by: Keyou Su
Resolution: Colors managed by theme tokens; dark/light switching supported during development; prototype reserves entry in settings page first.
(3) Gray small text too small; hope for font adjustment
Reported by: Keyou Su
Resolution: Improve contrast of gray small text and provide font size options; minimum font sizes for body and auxiliary text written into design specifications.
Developer Response:
"The prototype was converted from frontend pages and has some bugs; we can ensure these issues are resolved during delivery."
All issues above are recorded in the issue list and will be regressed item by item in the Alpha phase; fix results will be updated synchronously to the course blog.
5. Team discussion photos (Non-posed)
The photos below were taken during real prototype review and walkthrough sessions (not staged). They are also published in the course blog and used in slide 20 of the defense PPT.
Photo 1 (Dorm): checking the Modao prototype page against the running screen — non-staged.

Photo 2 (Classroom): walking through hotzone paths and page flows on a tablet — non-staged.

6. PPT Link
https://github.com/2310209788-del/cs2-intelligent-review-platform/tree/main/ppt
相关推荐
国央企创新负责人如何通过平台实现科研成果的快速识别与转化?.docx
科易网基于AI模型+行业大数据,在技术转移、成果转化、技术经纪、知识产权、产业创新、科技招商等垂直领域的多样化应用场景,推出科技创新领域的AI+工具与数智化解决方案,推动科技创新与产业创新智能化发展。
PotPlayer自用皮肤
自用 PotPlayer 皮肤,仅微调了字体大小与颜色,是按个人习惯调整,未做适配性调整。
2-shell及常用命令-续.pdf
2-shell及常用命令-续.pdf
【电力负荷预测】项目介绍 MATLAB实现基于BO-Transformer-LSTM贝叶斯优化算法(BO)优化Transformer-LSTM组合模型进行居民用户多步电力负荷预测(含模型描述及部分示例
内容概要:本文介绍了一个基于BO-Transformer-LSTM的组合模型,用于实现居民用户多步电力负荷预测的完整项目实例。该模型融合Transformer的全局注意力机制与LSTM的局部时序记忆能力,构建能够捕捉长时依赖与周期性特征的深度学习架构,并引入贝叶斯优化(BO)自动搜索最优超参数组合,提升预测精度与训练效率。项目涵盖从原始数据清洗、特征工程、滑动窗口样本构建、模型设计到超参数优化、反归一化及多步误差评估的全流程,强调时间顺序划分以防止信息泄露,并通过周期编码、外生变量融合等方式增强数据表达能力。文中提供了详细的MATLAB代码示例,包括数据预处理、网络搭建、贝叶斯优化配置和预测结果评估等关键步骤。 适合人群:具备一定机器学习和时间序列分析基础,熟悉MATLAB编程环境,从事电力系统预测、智能电网、负荷管理等相关领域的研究人员、工程师及研究生。 使用场景及目标:①应用于居民级或多用户聚合负荷的短期至中期多步预测(如未来24小时);②服务于配电系统运行调度、需求响应基线设定、家庭能源管理系统优化、分布式能源协调控制等实际工程场景;③为科研提供可复现的深度学习+贝叶斯优化的技术路线参考。 阅读建议:此资源以实际项目为导向,建议读者结合提供的代码逐步实现并调试模型,重点关注数据质量控制、时间特征编码、组合网络结构设计与贝叶斯优化稳定性之间的协同关系,在实践中深入理解多步预测中的误差累积机制与超参数敏感性问题。
【无人机航测】MATLAB重叠率与地面分辨率协同设计:航线间距及飞行方向优化系统 MATLAB 无人机航测重叠率设计|地面分辨率与航线间距计算
内容概要:本文档介绍了基于MATLAB R2025b开发的无人机航测重叠率与航线间距计算系统,旨在通过建立统一的航测几何计算体系,实现重叠率、地面分辨率与航线间距的动态协同设计。系统结合相机参数、飞行参数、地形数据与风场信息,通过自动化候选航向搜索与多目标优化,选择兼顾覆盖质量、飞行效率与安全性的最优航线方向。项目涵盖从参数输入、摄影测量计算、地形与安全修正,到候选路线生成、综合评价与结果可视化的全流程,并构建了五万组模拟数据用于算法验证与模型训练,支持桌面交互、批量处理与服务化部署,适用于多种复杂场景下的无人机航测任务规划。 适合人群:具备一定MATLAB编程基础和摄影测量知识的科研人员、工程技术人员及高校师生,尤其适合从事无人机航测、地理信息系统(GIS)、遥感与测绘等相关领域的从业者。 使用场景及目标:① 实现无人机航测任务中地面分辨率与航线间距的精确计算与动态优化;② 解决复杂地形与风场环境下航线方向选择困难、覆盖质量难以保障的问题;③ 通过自动化设计减少人为经验依赖,提升航测任务的可靠性与效率;④ 为后续的三维建模、正射影像生成与灾害应急评估等应用提供高质量的数据基础。 阅读建议:建议读者结合文中提供的完整代码示例与GUI界面设计,逐步实践参数配置、模型计算与结果分析流程。重点关注摄影测量公式推导、地形修正策略与多目标优化算法的实现细节,并利用模拟数据进行敏感性分析与模型验证,以深入理解系统的设计理念与工程应用价值。


