Discovery fatigue发现成本高
Researchers repeatedly scanned broad feeds to find a small number of relevant papers.研究者需要反复浏览大量内容,才能找到少数真正相关的论文。
A calmer way for scientists to discover, read and subscribe to academic papers让科研人员更高效地发现、阅读并订阅学术论文

Researchers browse high-quality sources such as arXiv every day, but discovering relevant work, managing interests and receiving updates were fragmented and difficult to control. Paper System brings browsing, reading, search and topic subscriptions into one coherent experience.
科研人员每天都会浏览 arXiv 等高质量论文来源,但发现相关内容、管理兴趣与接收更新的体验彼此割裂,也缺乏控制。Paper System 将浏览、阅读、搜索与话题订阅整合为统一体验。
Researchers repeatedly scanned broad feeds to find a small number of relevant papers.研究者需要反复浏览大量内容,才能找到少数真正相关的论文。
Tags and filters were difficult to create, understand, recover and maintain.标签与筛选条件难以创建、理解、恢复和持续维护。
Unpredictable email volume made subscriptions less useful over time.不可预测的邮件数量让订阅逐渐失去价值。
The study combined interviews, behavioral framing and journey mapping to move from an abstract idea of “paper discovery” to concrete moments of frustration and opportunity.研究结合访谈、行为还原与旅程地图,把抽象的“论文发现”转化为具体的挫折时刻与设计机会。
Use a first-person story to make fragmented routines visible to the team.用第一人称故事让团队看见分散的真实工作习惯。
Capture motivations, pain points, reading habits and expectations.整理研究动机、痛点、阅读习惯与期待。
Connect actions, touchpoints, emotions and ownership to prioritize opportunities.连接行为、触点、情绪和责任,以此确定机会优先级。



The core problem was not access to papers. It was maintaining a useful, predictable relationship with a changing field: researchers needed to see what was new, understand why it matched, and control when updates arrived.核心问题并不是能否访问论文,而是如何与不断变化的研究领域保持有用且可预测的关系:用户需要知道什么是新的、为什么与自己相关,以及更新何时到达。
ReadCube represented paper management, while Medium, Behance and PropertyGuru offered useful patterns for discovery, saved filters, detail pages and community signals.ReadCube 代表论文管理,Medium、Behance 与 PropertyGuru 则提供了内容发现、保存筛选、详情页与社区信号的成熟模式。
Balance popular, recent and subscribed content without making one feed do everything.平衡热门、最新与已订阅内容,避免用单一信息流承担所有任务。
Make tags and filters understandable, editable and recoverable over time.让标签与筛选条件长期可理解、可编辑、可恢复。
Keep citation data, abstracts, repositories and sharing actions close to the reading context.让引用信息、摘要、代码仓库与分享行为贴近阅读上下文。
Functional modules created a shared product boundary before detailed interaction work, reducing rework and helping product and design align on what the system needed to own.在进入细节交互前,功能模块先建立共同的产品边界,减少返工,也帮助产品与设计统一系统责任。
Paper System uses source data such as subject, author, content, publication time and categories, then turns it into four user-facing capabilities.Paper System 使用论文主题、作者、内容、发布时间与分类等基础数据,再将其转化为四类面向用户的能力。

Detailed interaction specifications covered feeds, empty states, advanced search, paper details, subscription editing, delivery schedules and email templates.详细交互覆盖信息流、空状态、高级搜索、论文详情、订阅编辑、接收周期与邮件模板。

Testing moved from a written plan and task script to moderated sessions, scoring, issue synthesis and a final prioritization discussion with product.测试从计划与任务脚本开始,经过主持式访谈、评分与问题归纳,最终与产品共同确定修复优先级。

Define goals, participant profile and realistic scenarios before sessions.在测试前明确目标、参与者画像与真实任务场景。
Record notes, task outcomes and satisfaction without over-interpreting one comment.记录访谈、任务结果与满意度,避免过度解读单一反馈。
Translate evidence into issues and align final priorities with product.将证据转化为问题,并与产品共同确定最终优先级。
Browse popular, recent and subscribed papers with advanced search when needed.浏览热门、最新与已订阅论文,并在需要时进入高级搜索。
Keep abstracts, authors, metadata and repositories in the reading context.在阅读上下文中保留摘要、作者、元数据与代码仓库。
Manage topics and choose when updates arrive instead of accepting a noisy default.管理兴趣主题并自主决定更新频率,而不是被动接受默认推送。
For knowledge products, relevance is not enough. Users also need to understand and control the system that decides what reaches them.对知识型产品而言,相关性还不够;用户还需要理解并控制内容为何、何时到达自己。
Paper System strengthened my practice of connecting research evidence, information architecture, interaction detail and usability testing into one product decision loop.Paper System 让我进一步建立起一条完整方法:把研究证据、信息架构、交互细节与可用性测试连接成同一个产品决策闭环。