Signal to BuildSIGNAL TO BUILD中文Owner sign in ↗Explore curated knowledge, course notes and methods, with their existing content and sources preserved.
This index lists complete English editions. Additional Chinese notes and bilingual updates remain available in the Chinese index.
Understand the evolution of the newtype knowledge system through long-term notes, source-based Q&A, and Agent execution; front-load the import boundaries of Obsidian and NotebookLM and the reading entry points, while preserving the original architecture, tools, and historical viewpoints.
Huang Yihe 2024-03 → 2026-05 Full evolution of model evaluation/fine-tuning/hardware real-world testing. 505 posts, covering the evolution of GPT/Claude/Gemini/Grok/DeepSeek/Qwen/Llama/Kimi/GLM/Minimax, Mac cluster benchmarks, Unsloth GRPO fine-tuning, Scaling Law debates, output length bottlenecks, TPU vs GPU, the strategy of general base models plus upper-layer vertical fine-tuning, model routing, and model committees.
30+ selected Q&A entries from Huang Yihe (newtype)'s Planet community from January to July 2025. Beginner learning path (Learning by doing: try all three models → ChatWise/Cherry Studio → OpenRouter) / Network conditions (US residential IP / Claude Code subscription vs API) / Multi-platform strategy (Domestic: WeChat Channels 68k + Bilibili & Douyin 40k combined + Xiaohongshu 10k / Overseas: YouTube + X + Medium + Reddit + newtype.pro) / Three reasons for not joining WeChat groups (filtering / education stage / time and energy) / AI Coding tool stack (Cursor + Claude Code complement each other) / Reasons for being bearish on Workflow vs bullish on MCP / Prompt House design philosophy (2-step MCP tool call) / Video production workflow (A Roll + B Roll + Screen Studio + OBS) / Three reasons for not subscribing to ChatGPT / RAG vs Fine-tune decision-making / Note structure Input-Processing-Output / Three things that don't rely on self-discipline (sense of mission + physical stamina + not forcing yourself). This is a "Q&A scenario" supplement to [[newtype-mindset-super-individual]] and [[newtype-knowledge-system]]
4 high-quality Prompt templates + 1 CLAUDE.md sample curated by newtype (Huang Yihe). YouTube Podcast Deep Comprehension for Gemini v2 (5-step method: chapter extraction → retelling → bilingual → topic background → learning assistance) / Social Media Deconstruction (5 frameworks: Core Insight / Structure / Psychology / Virality / Takeaways) / Creative Divergence Mode (4 lenses: Gap Analysis / Reframing / Cross-Domain / Breakthrough) / Weekly Synthesis Report (Markdown folder analysis → cross-document synthesis) / CLAUDE.md MCP configuration sample (Tavily + Sequential Thinking + Prompt House + Super Analyst Pro). This is a supplement to the "mature prompt library" of [[newtype-prompt-engineering]]
A complete breakdown of 3 representative works in newtype's Skill system. Super Analyst v1 (intelligent selector for 12 analysis frameworks + integration with prompt-house-local MCP) / Super Analyst Pro 2.0 (7-Stage Workflow + dual Sequential Thinking + Bilingual Web Search + 3-level complexity adaptation) / Super Writer (Material Collection + 7 Style Analysis methods + 6 Writing Methodologies + triple Sequential Thinking). Each Skill is an engineering paradigm of "Sequential Thinking + MCP + framework library". It is the practical proof of the judgment "Skill-ification is the end state" in [[newtype-mcp]]
Complete documentation for two core CLI tools built by newtype. The Oh My OpenCode customization guide (1,729 lines) covers two approaches (configuration files versus forking the repository), Chinese-localized names for 10 agents (for example, Sisyphus → Xixifusi and oracle → Shenyuzhe), automated Perl replacement scripts, GitHub Release versus npm deployment, the SUL-1.0 license, and an eight-step walkthrough. The newtype CLI multi-agent content team covers the nt binary, eight specialist AI agents (researcher/writer/editor/fact-checker/extractor/analyst/archivist/chief), eight core commands (research/write/edit/fact-check/analyze/extract/archive/pipeline), JSON envelopes and NDJSON streams, four exit codes, and 12 analysis frameworks. This is the OpenCode-side tool stack complementing [[newtype-claude-code-best-practices]].
5 curated essential articles from the Nielsen Norman Group accessibility / barrier-free design knowledge base. Four core themes: ① Boundaries of three easily confused concepts — Accessibility (serves only users with disabilities, narrow and evaluable, anchored to WCAG standards) vs Universal Design (one design serves all people without variants, mostly used for physical environments) vs Inclusive Design (allows multiple design variants to cover differences in age, economy, region, language, race, ability, etc., the first choice for digital products because digital interface modification costs are low), with positive examples including Tinder's 30+ gender options, Pinterest's hairstyle filters, Sephora's age group filters, diverse illustrations from Airbnb/Atlassian, and unrestricted surname fields; ② Visual science of Dark Mode — Light mode (dark text on light background) outperforms across all dimensions of visual acuity, proofreading, and quick nighttime browsing for users with normal vision (2013 Ergonomics / 2017 MIT AgeLab / Human Factors: the smaller the font size, the greater the linear advantage of light mode); dark mode enables faster reading for users with cloudy ocular media such as cataracts, and saves 67% power on OLED screens at 100% brightness (only 14% power saving at 30% brightness, Purdue); long-term light mode reading is associated with myopia / choroidal thinning (2018 Nature Scientific Reports); default to light mode + provide a dark mode option + automatically follow system settings; ③ 8 major design pitfalls of dark mode — graphic transparency (use SVG/WEBP/PNG instead of JPG), depth created by shadows fails (switch to darkest color for base layer, lightest color for top layer), fonts that are too thin/too thick/insufficient contrast (WCAG minimum 4.5:1 for normal text, Coinbase's blue button at 4.59:1 barely meets compliance but is still hard to distinguish), high-saturation colors become invisible, inconsistency between dark apps and light web pages across channels, dividers disappear, floating components blend into the background, inverted QR codes cannot be scanned (foreground and background keep the same color across both modes); when to prioritize dark mode (long-duration use / frequent use / low-light environments / media-light content); ④ User research methods — word-of-mouth recruitment for Screen Reader users (contact blind organizations such as NFB to build user panels), in-person is better than remote, print research plans, dual-channel recording of five elements via document camera + screen recording, audio feedback for time announcements; as well as 16 usability guidelines for IVR telephone voice tree systems (auditory information accounts for only 11% of human information intake vs 83% for vision, IRS telephone trees reach as deep as 10 layers). Audience = product managers, terms explained in industry language, decision-oriented (when to use / how to choose / counterexamples).
20 curated articles from the 'Artificial Intelligence/AI' topic at Nielsen Norman Group, under the 'AI Product and Conversational Interface Design' subset (focused on the product and interface layer, distinct from the research practice layer). Five core axes: ① Conversational AI experience — 6 types of conversations with generative AI (Search Query / Funneling / Exploring / Chiseling / Pinpointing / Expanding; conversation length ≠ success), Less Chat More Answer (cut pleasantries / scannable lists / answer first then details via the 'Truncated Pyramid' / say plainly when you can't do it / specific over generic), Explainable AI in chat (three placement methods for Source Citations: Perplexity inline / Copilot sidebar / can hallucinate sources; Step-by-step reasoning; clear actionable vs vague Disclaimers; avoid anthropomorphic wording), Sycophancy (contradicting facts to please users · human feedback reward loopholes · recanting when asked 'Are you sure?' · three mitigation approaches). ② Anthropomorphism and trust — 4 degrees of anthropomorphism (Courtesy / Reinforcement / Relayed roleplay / Companionship), Smarts over Sentience (perceived intelligence ↑ trust ↑ · perceived emotion ↑ trust ↓ · warm AI has a 10–30% higher error rate · professionalism ≠ coldness), early user behaviors of Accordion editing and Apple Picking. ③ Prompt UX — 3 types of prompt suggestions (Use-Case / Autocomplete / Followup), design of use-case prompt suggestions (simple pills vs complex cards/carousels · two stages before and after login · specific over vague · personalization), CARE prompt framework (Context / Ask / Rules / Examples), 4 uses of Prompt Controls (discoverability / education and inspiration / constraint / promoting follow-ups · 50% include follow-ups, 77% are multi-turn · 4 best practices). ④ Generative UI (GenUI) — GenUI vs Vibe Coding vs AI-assisted design (who decides to build this UI · execution failure vs judgment failure), the most exciting GenUI development is the return of buttons and checkboxes (AskUserQuestion to collect context), Promptframe (documenting wireframes + AI prompts · 6-step process · Blue Apron case study), 5 fixes for designing interfaces with prompts (precise visual keywords / lightweight reference images / analyze first then generate / real mock data / attach code). ⑤ AI assistant positioning — 4 roles of AI as a UX assistant (content editor / research assistant / creative partner / design assistant · 841-person survey, 92% already use it), AI as a creative partner (P&G 776-person experiment · individuals using AI outperform groups without AI · the Top 10% are still groups), initial review of ChatGPT Agent booking lunch (96 sources in 6 minutes · 11 minutes to fill forms · human-in-the-loop · decision gate), Google AI Mode evaluation (Query Fan-out + RAG · 7 users · poor discoverability / naming confusion / overly long answers), why users switch between GenAI and search (6 out of 9 users switch back and forth · AI clarifies questions, search verifies facts · even citations struggle to build trust). Audience = non-technical product managers, terms explained inline, decision-oriented (when to use / how to choose / counterexamples).
22 curated pieces from the Nielsen Norman Group analytics and metrics knowledge base. Centered on the core theme of "how to use data to prove the value of UX", it clarifies four key topics: ① Choosing the right metrics (avoid Vanity Metrics — isolated numbers that "only go up and never down" such as total downloads, total page views, and total user count; replace them with ratios/proportions, time frames, user segment breakdowns, and the DAU/MAU Stickiness Ratio. Conversion Rate = the most important metric, but it is affected by external interference from price, seasonality, and marketing. Macro vs Micro Conversions = process milestones plus secondary actions. Success Rate (for usability testing, when user intent is unclear) vs Completion Rate (for analytics tools, when user intent is clear). For time on page/page views, "more ≠ better"; you must distinguish between productive tasks and engaging tasks); ② Division of labor between two quantitative methods (Analytics answers "what users did" — large scale, low cost, no context; Quantitative Usability Testing answers "whether users can complete tasks" — ~40 participants, expensive, supports causal inference; the two are complementary); ③ How to calculate UX ROI (four steps: collect UX metrics → select KPIs → convert to monetary value → report responsibly. Health insurance case: 23,000 → 1,100 tickets × $6 = $131,400 saved per month / $1,576,800 saved per year / $4,655,400 return over three years; design system case: calculating just one video plugin was enough to persuade stakeholders; three major misconceptions = ROI is not only about money / does not have to be exact / does not cover every detail; Don't Overthink); ④ Statistical rigor (Statistical Significance vs Practical Significance, example of 72s vs 75s; Confidence Intervals at 95%, e.g., 4.3min → 3.6–4.9min, with sample size and consistency determining interval width; decision-making when results are not significant = 60% maximum likelihood probability plus qualitative insights; Benchmarking requires ≥20 participants, conducted annually, to quantify business value). Supplemented by content on Confounding Variables (randomization, within-subjects design, age, seasonality, pandemics, competition, experience, prior opinions), Convenience Sampling vs Probability Sampling (quantitative studies ≥40, surveys ≥100, simple random vs Stratified), Triangulation (cross-checking multiple methods/metrics to improve credibility), Statistically-Generated Personas (qualitative foundation + ≥100 participants, ideally 500+, latent class analysis), and Long-Tail Data (power-law distribution complemented by the Pareto principle). Intended audience: product managers.
20 curated essential pieces from the Nielsen Norman Group knowledge base on content strategy and web writing, organized around three core axes: ① Content governance and process — Content Strategy is the high-level plan that precedes tactics (four phases: Plan, Create, Maintain, Unpublish); Content Strategy vs UX Writing (the former governs processes while the latter writes copy, with a 9-dimension division of responsibilities), Content Design, Content Inventory (listing all assets) vs Content Audit (assessing quality; inventory first, then audit, with audits conducted at least every 3 months), integrating Content Standards into design systems (global + specific, reviewed annually), 3 intranet content management models (centralized/decentralized/hybrid), 5 types of content user stories in Agile (including content research Spikes), and Information Architecture (IA) vs Sitemap. ② How to write content people will actually read — People primarily read online via Scanning (13 years, 500+ participants, 750 hours of eyetracking data); 5 classic scanning patterns (F-pattern/Spotted/Layer-Cake/Commitment/Lawn Mower) plus new patterns (Pinball: 158 jumps in 14 seconds; Love-at-First-Sight SERP: 20% of users view only one result); the Inverted Pyramid with conclusions placed first; 7 methods for concise writing (including Kill Your Darlings); 5 layout approaches for long-form content (summaries/bold highlighting ≤30%/key points/callouts/visuals); microcopy (<3 sentences) with the 3I framework (Inform/Influence/Interact) + 3C framework (Clarity/Concision/Character); 4 dimensions of Tone of Voice (funny-serious/formal-informal/respectful-disrespectful/enthusiastic-matter-of-fact). ③ How to validate — Content testing uses open-ended information-finding tasks + open-ended follow-up questions; Eyetracking setup (fixation maps: 8–12 participants; heatmaps: 39 participants; Tobii Spectrum). Target readers = product managers; jargon is explained in-line; decision-oriented (when to use/how to choose/counterexamples).
29 curated pieces from the Nielsen Norman Group customer journey knowledge base. The core thread = visualization and operation of "the holistic experience of a user interacting with a product/service across time and channels to achieve a goal". Core artifacts: Journey Map (five elements = Actor / Scenario + Expectations / Phases / Actions-Mindsets-Emotions / Opportunities; top-middle-bottom three-section structure; Emotion Curve), Experience Map (product-agnostic baseline), Service Blueprint (enterprise perspective with four swimlanes = Customer Actions / Frontstage / Backstage / Support Processes; three dividing lines = Line of Interaction / Line of Visibility / Line of Internal Interaction; 1982 Lynn Shostack), Empathy Map (four quadrants: Says/Thinks/Feels/Does), User Story Map (three layers of Activities-Steps-Details + release line, Jeff Patton, Lean UX conversion to Epic/User Story), User Flow (step path for a single product and single task, linear flow diagram), Ecosystem Map, Process Map, Sankey Diagram (nodes-links-columns, used for path analysis, Mixpanel/Amplitude/GA). Method decisions: As-is vs To-be, Assumption-First vs Research-First, low-fidelity vs high-fidelity, hybrid three-step method. Seven journey map analysis methods (unmet expectations / redundant touchpoints / emotional lows / cross-channel breakpoints / time consumption / Moment of Truth / highs). Project operations: journey mapping impact survey (300+/259 respondents, 64% collaborative creation, 56% digital tools, 7 outcome success metrics, 3.5 barrier lines, 8-factor ranking with customer participation highest at 6.08, visual design lowest at 2.18), 7-step interactive Figma maps, CX transformation framework (4 domains x 4 building blocks = 16 tasks), Journey-Centric Design vs Journey Management party hat model, UX vs Service Design, 5-step service blueprint and practice survey (97 respondents, 56% deliverables / 20% frameworks / 15% collaboration tools, focused type <2 weeks + 1 week, panoramic type <1 month). Research foundations: Contextual CUEs framework (Culture / Unspoken / Environment, Distributed Cognition DCog), User Panel 101 (customer-type/goal-type, B2C/B2B/two-sided marketplaces, 6-step setup). Practical cases: four types of luxury shoppers (browsers only / occasional splurgers / big spenders / stylists, consideration stage most critical), healthcare digital interactions (9 participants, 93 diary entries of interactions, 78 on smartphones, 32 on portals, 2/3 digital pain points), four stages of AI image generation (define-explore-refine-export, Midjourney, inspiration-oriented vs deliverable-oriented), UX storytelling (story triangle / Freytag's five-act story pyramid). Target readers = product managers, terms explained in-line, decision-oriented.
17 curated articles from Nielsen Norman Group’s “Organization and Design Process” collection, covering “what actions you take and what you produce during the design process.” Five major themes: ① Discovery Phase — Discovery Mindset (focus on problems rather than solutions, counter Anchoring effect + Confirmation Bias, 3 tactics: Reframe the Goal/Highlight Unknowns and Assumptions/Icebox Favorite Ideas), 7 tips for successful discovery (secure buy-in first/clarify roles and norms/write a problem statement that does not imply solutions/cluster and vote on unknown priorities/timebox to counter planning fallacy, 7-week sample plan: weeks 1–2 framing + weeks 2–4 research + weeks 3–5 analysis + weeks 6–7 opportunities and ideation/all team members participate in research/collaborative analysis and synthesis with How-Might-We), CSD Matrix (Certainties “We know”/Suppositions “We believe” testable/Doubts to be researched, rows grouped by user·business·technology·people·process·status·priority, goal = convert doubts and suppositions into certainties), 4 steps to kick off a new UX project (enter kickoff with a question list/expose information gaps through verbal discussion/schedule based on gaps/pave the way for the future, adapted for Agile + Waterfall); ② Wireframes and Collaborative Sketching — 4 steps to draw wireframes (set aspect ratio 1024x768/1920x1080·navigation and search·largest elements: headings in thick lines, body text in thin lines, images marked with crosses·details: dropdowns, radio buttons, buttons), magic formula to get stakeholders to sketch (thick Sharpie markers + small 8-grid paper + time limit of 5 minutes per sketch + intentionally ugly examples = relaxed sketchers), Promptframes (wireframes + generative AI prompts, replacing lorem ipsum placeholder text, 6 steps: define context, role, organization, goal, tone + write recording prompts + run + collaborate to test and optimize + iterate + manual refinement, 2–3 variants, for internal use only, not for senior leadership); ③ Prototype Fidelity — 7 decision questions for low-fidelity vs. high-fidelity, clickable vs. static prototypes (can all responses be predefined/can it be trialed multiple times/can all issues be fixed/is the design stable/is someone acting as the computer/is page flow the focus/do dynamic changes need to be observed), why test prototypes (save cost/avoid risk/iterate); ④ Design Specs and Deliverables — Dual design spec package (Figma design file: interaction flows·visuals·layout grids·components·real content·accessibility tab order and alt text; Linear development task ticket: goal·scope including inclusions and exclusions·assumptions·success metrics·Figma link·notes) + 4 tips for alignment with developers (invite developers early/leave time for comments/prepare tradeoffs/record and sync updates), 50+ deliverable glossary terms (affinity diagram/antipersona/archetype/CSD matrix/empathy map/experience map/Kano model/MoSCoW/RICE/RACI/service blueprint/persona including proto/qualitative/statistical/prototype/paper prototype/promptframe/sketch test...), most commonly used deliverables (survey of 86 practitioners: static wireframes/interactive prototypes 67%/flowcharts/site maps/usability reports; select by audience: management views interactive prototypes + usability reports + journey maps·clients view interactive prototypes 67% + high-fidelity mockups 47% + competitive analysis 34%·developers view interactive prototypes 63% + flowcharts 55% + style guides 51%); ⑤ Expert Review/Workshops/Storyboards/Surveys — Expert Review (three types: heuristic evaluation/independent design critique/expert review, external perspective is more objective + uncovers blind spots, 5 components: strengths list·issues list·severity high/medium/low·improvement suggestions·best practices, complementary to usability testing, conduct regular reviews every 2–5 years), Workshop Facilitation (4 goals: full and equal participation Diverge-Converge/promote mutual understanding/inclusive collaborative decision-making/shared responsibility, 6 principles: listen·inclusive space·welcome improvisation·authentic and transparent·do not give advice·embrace constructive conflict, toolkit = materials + activities post-up/affinity diagram/forced ranking/role-play/Playback + techniques silence/balancing/connecting), Storyboard (three elements: scenario + visuals + captions, differences from journey maps, 6 steps, low-fidelity sticky notes can be rearranged), In-Process Surveys (4 stages of the design cycle: discover/explore/test/listen, corresponding types for each stage: discovery phase qualitative surveys + stakeholder surveys + diary studies/exploration phase statistical persona surveys + competitor customer surveys/testing phase post-task SEQ single ease question + post-test SUS System Usability Scale 10 items/listening phase NPS 0–10·CSAT 1–5·CES, counterexample = mistakenly thinking surveys are fast and simple). Audience = non-technical product managers, terms explained in-line, decision-oriented (when to use/how to choose/counterexamples), all numbers retained.
19 curated articles from Nielsen Norman Group’s “Visual Design” topic, covering the design systems and icons subset. Six clusters: ① Design systems and style guides (Design System = parent, containing Component Library / Pattern Library / Style Guide; Style Guide = child, divided into content, brand, and frontend types; three implementation approaches with three-way tradeoffs among cost, customization, and brand consistency: adopt / adapt / create; examples include NASA 1976, Material, Carbon, USWDS). ② Icons and signifiers (Icon usability = Recognition, whether users recognize the shape, + Interpretation, whether users understand the meaning; 4 quality criteria: Findability / Recognition / Information Scent / Attractiveness; isolated vs. contextual testing matrix; Accordion signifier experiment with 11 prototypes and 5 variants → Caret dropdown indicator performs best, right arrow ≈ no icon; floppy disk = save 83%, circular arrow = sync/refresh 73%, down arrow + container = download 78%; Sparkles icon AI overuse · none of 107 people associated it with AI · 16.82% misread it as bookmark/favorites). ③ Visual style evolution (Skeuomorphism as learning aid vs. visual trend → Flat → Neumorphism low-contrast pain point; Glassmorphism since 2020 with transparency + background blur + outlined gradients · Apple SwiftUI / Microsoft Fluent Acrylic / Vision Pro · WCAG contrast risk). ④ Clickable element identification (Beyond Blue Links 5 types of cues · Interaction Cost) + Banner Blindness (selective attention · proximity poisoning · hot potato effect · pseudo-advertising · persistent since 1997). ⑤ Animation and motion (6 uses: feedback / state transitions / spatial metaphor navigation / signifiers / attracting attention / attention hijacking dark pattern; restraint principle). ⑥ Visual testing (Attitudinal vs. Behavioral; 5-second test / first click / preference test / open and closed word selection; Microsoft Desirability Toolkit 118 word cards 2002 controlled vocabulary · trimmed to under 25 words · Venn diagram grouping; Mood Board 4–5 keywords → visual collage). Also includes brief summaries and attribution for 3 research/team method misfit articles (affinity diagramming / UX mapping visuals / groupthink / NPS shortcomings). Audience = product managers; in-line terminology explanations, decision-oriented (when to use / how to choose / counterexamples), all numbers retained.
15 curated articles from the Nielsen Norman Group e-commerce UX knowledge base. Reorganized into 8 sections along the shopping decision journey: ① Shopper typology and decision models — 5 types of general e-commerce shoppers (product-focused / browsers / researchers / bargain hunters / one-time) + 4 types of luxury shoppers (window shoppers / stylists / occasional splurgers / big spenders), compensatory vs noncompensatory decision-making (use facets to filter out options → use comparison tables to weigh tradeoffs; 5–7 options is the tipping point); ② Product Listing Page (PLP) browsing — traditional pagination vs Infinite Loading (only use when <40 items and strong filtering is applied) vs Show More button, must display total item count + loading progress (Lululemon: 40 of 333 items) + Save the Spot to preserve browsing position; ③ Category Page — traditional vs Hybrid (place educational content at the top of the listing page instead of forcing a detour), Bonobos direct access vs Eddie Bauer detour; ④ 6 rules for product photos — Large / Consistent / Specific / Dynamic hover / Contextual / Inclusive; ⑤ Multi-variant display — simple variants (color, size) use on-page attribute selectors vs complex variants (TV product lines) use separate pages + similar products module, Best Buy product line dropdown as a counterexample vs Costco separate pages; ⑥ Comparison Table — ≤5 items / static vs dynamic / sticky headers / highlight differences / convert columns to labels on mobile; ⑦ Trust and globalization — Diverse Models / model selection feature / international sizing with both local and original size / brand-specific size charts / one-click Translate All for UGC reviews / Socially Conscious practices (Toms donates $1 for every $3 earned) / AR try-on for makeup and furniture (accuracy determines usefulness); ⑧ Luxury e-commerce — the Consider stage is the weak point, scarcity, price range, deep personal connection. Target readers: product managers; jargon explained in-line, decision-oriented (when to use / how to choose / counterexamples).
6 curated articles from the Nielsen Norman Group marketing channel UX knowledge base, covering two core threads: ① Email — Two studies on emojis in subject lines (emojis increase negative sentiment by 26%, lead to slightly lower value ratings, and do not significantly improve open rates; they should only be used to supplement context or emotion, e.g. '🎓Congratulations on your admission'); Top 10 design mistakes in the unsubscribe experience (unclickable footer links on mobile, hard-to-find links, cumbersome processes, counterintuitive checkboxes, mandatory login, requesting feedback before unsubscribing, offensive tone, inducing resubscription, sending confirmation emails, delayed removal of up to 10 days; the ideal experience is one-click direct access + polite confirmation page + no extra steps; bad unsubscribe flows push users to mark emails as spam, harming deliverability); 16 guidelines for status trackers & progress updates (combining pull (active user query) and push (passive notification) works best: 12 guidelines for trackers + 4 for progress updates, use user-facing language instead of backend jargon such as 'label generated', pin the latest status to the top, allow frequency adjustments, support STOP to unsubscribe). ② Social Media — 6 types of enterprise interactions (Discover / Research / Engage / Purchase including social selling vs social commerce / Support / Promote); 3 social media UX insights (based on 96 users over 11 years + 23 user tests: don’t post too frequently — 5–8 posts per week as baseline, platform-specific breakdown: 2–6 posts per day on Twitter, up to 15–20 during campaigns; 1 post per day on IG/FB/TikTok; 6–8 ephemeral content posts per day; use interactive features like polls and quizzes, e.g. NBA poll with 5600 responses / Xbox naming poll with 71,000 votes; include clear CTAs such as 'Swipe up'); Deceptive dark patterns in K-Pop direct messaging apps Bubble and Universe (paid access creates a false one-on-one private chat illusion in one-to-many broadcasts, Kakao-style '1' read receipt marker, personalized username variables, gamified subscription timers and anniversary features to build sunk cost, 'broken heart' guilt manipulation when unsubscribing). Target audience: product managers; terms are explained inline, decision-oriented (when to use / how to choose / counterexamples).
16 curated articles from the Nielsen Norman Group Human-Computer Interaction (HCI) knowledge base. Coverage spans three major sections: ① Hard laws of motor control and cognition — Fitts's Law (time to reach a target T depends on distance D and target width W; larger targets are faster · screen edges = infinite targets · padded targets · Woodworth two-phase movement model · distance comparison of linear vs rectangular vs pie menus · place buttons close to the prior action), Accot-Zhai Steering Law (movement time within a bounded path T=a+b·A/W; narrower and longer paths are harder · L-shaped corridors in hierarchical menus · vector triangle buffer zones · Mega Menus · forgiving paths), Touch Targets (minimum 1cm×1cm / 0.4in · average finger width 1.6–2cm · thumb 2.5cm · spacing to prevent accidental taps · View-Tap Asymmetry · use larger 2cm targets for children/older adults/one-handed use), design process (Iterative design +38% per round · Parallel design merge +70%, then iterate for 152% · Competitive testing). ② Interaction patterns and interfaces — State-Switch controls (dual control: state indicator + toggle button vs single control · Mute button counterexample WebEx · positive example Zoom), Tutorials vs Contextual Help (push vs pull · short-term memory load · progressive disclosure), Task Analysis (task vs goal · contextual inquiry · HTA hierarchical task analysis diagram · GOMS · cognitive complexity), four core Data Tables tasks (lookup/compare/edit single row/batch operations · freeze header and first column · zebra striping · in-place editing vs non-modal side panel). ③ Smart homes and new scenarios — 7 best practices for smart devices, 5 guidelines for onboarding and connection, 5 motivations for smart homes (convenience/peace of mind/saving money/useful data/ambience), 3 user types (Managers/Everyday Users/Restricted Users) + Zone×Role dual axis · agentic execution · invisibility of automation · ability to degrade to Dumb Work, 7 principles of smart notifications (timely/relevant/specific/intensity/frequency/channel/adjustable) × 3 types (Reactive/Proactive/Optimization), Smarts over Sentience (warm AI increases error rate by 10–30% · RAG), 13 practices for Virtual Queues, 10 guidelines for AR calibration. Audience = product managers; terms explained in-line, decision-oriented (when to use/how to choose/counterexamples).
6 essential articles from the Nielsen Norman Group Information Architecture (IA) knowledge base. It starts with a clear definition of IA: IA is the practice of organizing, categorizing, and labeling content and designing relationships between content (it is not a single deliverable, but an ongoing optimization process). There are three core axes: ① The three models of IA — Visible Navigation (Front Stage, including menus/links/breadcrumbs/accordions, aligned with users' mental models) / Underlying IA Structure (Back Stage, the blueprint of full-site page relationships ≈ Sitemap) / Taxonomies & Metadata (concept relationship maps that organize 'concepts' rather than 'content', more technical); it also clarifies IA vs sitemap (IA is the overall practice including Content Inventory, Content Audit, IA UX research such as Card Sorting/Tree Testing, and taxonomy labels, while a sitemap is only its visual output, with nngroup.com as an example using 3 levels of blue/green/yellow nodes). ② Structural choices — Flat vs Deep Hierarchies: flat structures offer strong visibility and low interaction cost but cause information overload; deep structures keep the top level concise but increase cognitive load and lengthen click paths; examples include the New Zealand government's flat structure with 15 top-level categories vs Ernst & Young's deep structure with 5 top-level categories (11 subcategories under Services, 10 further subdivisions under Consulting); the conclusion is to avoid extremes, moderately flat structures outperform moderately deep structures, and user research is key. ③ Navigation and content page implementation — Left-Side Vertical Navigation (accommodates multiple categories, is extensible, easy to scan, users devote 80% of their attention to the left half of the screen; Abel Restaurant uses 13 categories, Slack/Gmail also use it; counterexamples include Accenture hiding everything under a single 'Services' item, Nua Bicycles with a content-to-chrome ratio changing from 5:1 to 12:1; 5 best practices: prominent left alignment / no duplicate navigation / icons paired with text / long menu optimization / mobile adaptation); full Taxonomy workflow (Controlled Vocabulary, hierarchical, non-extensible, supports Faceted Navigation as used by the US Library of Congress and search suggestions as used by the World Bank; compared with Thesaurus and Ontology, 8-step construction method at BARTOC.org); 5 common mistakes on privacy policy/terms of service pages (obscure language, lack of overview, poor typography with font size below 14pt, incomplete navigation, placement not matching user expectations; positive examples include LinkedIn/eBay/Google, counterexamples include Forever21/Ameren/Spotify using all caps). Target audience: product managers, with in-line explanations of jargon, decision-oriented (when to use/how to choose/counterexamples).
Nielsen Norman Group 'Design Patterns' ethics, globalization, and emerging technology subset: 15 curated articles, reorganized by theme into five clusters. ① Design Ethics and Deceptive Patterns (Dark Patterns, coined by Harry Brignull in 2010): 5 categories of tactics (Obstruction · Visual/Text Traps · Nagging · Emotional Manipulation · Sneaked Preselection) + Sludge + the line between persuasion and deception (LinkedIn/Booking.com) + 9-question cognitive walkthrough; three types of Sneaking (Forced Continuity · Hidden Costs · Sneak into Basket, with negative examples from Spotify/Ticketmaster/Airbnb/GoDaddy/LinkedIn vs positive examples from Apple/United); Dangerous UX (high-risk actions placed adjacent to harmless actions, shape coding from Chapanis' B-17 landing gear case · visual signals · increasing distance per Fitts' Law); legal risks including GDPR/CPRA/Amazon Prime lawsuits. ② Inclusivity and Autonomy (boundaries between Inclusive Design ≠ Accessibility (WCAG) ≠ Universal Design; 5 pattern categories: senior-friendly fonts/dark mode · global surnames · 30+ gender identities · inclusive filtering · diverse illustrations; three methods for user Autonomy: Customization · Scannability · Timing and Sequence, with the cattle chute metaphor + meaningful defaults). ③ Global Cross-Cultural Design (Translation vs Localization; Hofstede's 6 dimensions: Power Distance/Individualism vs Collectivism/Masculinity vs Femininity/Uncertainty Avoidance/Long-term vs Short-term Orientation/Indulgence vs Restraint; 6 tips for language switchers · language vs country switchers · local language names Español/Deutsch; international sizing with local size + original size + body measurements + brand-specific charts; WeChat/Alipay login; 94% of TV viewers use a mobile phone as a second screen; Collaboration among the 5 elements of omnichannel coordination + QR codes to reduce interaction cost). ④ Emerging Interactions (applying 10 VR heuristics to Oculus Quest; 3-step AR testing with 11 users · discoverability · glasses conflicts · no gratuitous AR; 4 uses of GenAI Prompt Controls: discoverability/educational inspiration/setting limits/facilitating follow-up, icons must have labels with negative example of ERNIE Bot magic wand). ⑤ UX Copy and Recommendations (three tiers: Long ≥3 paragraphs/Short 2-3 paragraphs/Micro <3 sentences; 5 guidelines for recommended content: personalized top placement · source labeling · categorization · feedback mechanisms · real-time updates; Explicit vs Implicit Differentiation comparison table + progressive disclosure + SERP). Target audience = product managers, with in-line explanations of jargon, decision-oriented (when to use/how to choose/negative examples).
Essentials of 15 patterns from the "Feedback and Dialogs" subset of the Nielsen Norman Group design pattern knowledge base. Six major themes: ① Dialogs and interruption grading — Confirmation Dialogs (8 rules: only use for high-severity actions · do not use for routine actions · state consequences clearly · use descriptive buttons instead of Yes/No · progressive disclosure · do not default to Yes · require typing DELETE for high-risk actions · provide a "don't show again" option), Modal vs Nonmodal (5 drawbacks of modals that force response/obscure the background + 7 scenarios where they should/should not be used), 10 problematic popup trends Popups (before page load/right after login/asking for email/requesting feedback before any user action/during critical tasks/multiple in a row/redirecting to external sites/interrupting content/GDPR Cookie/promoting apps without stating benefits — with nonmodal/banner alternatives for each). ② Error messages — three categories of Error-Message guidelines (4 for visibility/4 for communication/4 for efficiency) + 4 principles for improving visibility (place close to the source · redundant for accessibility · grade by severity · do not show too early) + Hostile Patterns counterexamples (premature prompts · misuse of error styling). ③ Loading and progress — Progress Indicators (loop animation for 2–9 seconds/percentage bar for 10+ seconds · provide estimated time for >10 seconds · do not use static indicators · do not warn users not to click repeatedly), Skeleton Screens (three types: static/animated/wireframe · do not use for <1 second · comparison of the three), 13 practices for Virtual Queues (explain purpose · status tracker · auto-refresh · notifications · re-entry · non-intrusive entertainment). ④ Tips and micro-feedback — Tooltip (hover on desktop · do not place critical information · arrows · consistency), Info Tips (i/? icons · three major pitfalls: wasting time/hiding critical information/interrupting flow), Transactional Notifications (SMS vs push · 50–240 characters · only use for urgent matters · allow unsubscribing), 5 rules for User-Feedback Requests (ask after the task · do not spam emails · surveys <1 minute with 4–5 questions · flexible formats · thank and incentivize). ⑤ Empty States 3 goals (communicate status · learning prompts · shortcuts to key tasks). ⑥ Data presentation — Choosing Chart Types (3Cs: Context/Clutter-free/Contrast · bar/line/scatter · avoid pie, bubble, and Sankey charts) + 4 principles for reducing cognitive load in forms Few Guesses (structure/transparency/clarity/assistance · single column · do not use placeholders). Audience = product managers, in-line explanation of jargon, decision-oriented (when to use/how to choose/counterexamples), all numbers retained.
19 curated articles on the forms and input subset of Nielsen Norman Group design patterns, organized thematically into five sections: ① Overall forms and validation (10 form usability guidelines raise error-free first-submission rate from 42% to 78% · 10 error-reporting rules: inline validation/error messages placed next to fields/dual-channel color + icon cues/use modals sparingly/do not validate before input is complete/validation summary cannot be the only indicator/do not use tooltips for errors/offer extra help after 3 repeated errors · four date-input patterns: calendar/scrolling/split dropdown/direct input + internationalization ambiguity between 10/11 formats) ② Selection controls (three checkbox forms: standalone/list/nested tri-state · comparison of checkboxes vs radio buttons vs dropdowns vs toggles · 5/15 option threshold for listboxes vs dropdowns + Steering Law · parameter controls: sliders/knobs/2D matrices/linked coarse and fine adjustment + Accot-Zhai steering law · steppers for small adjustments with 1cm touch targets) ③ Toggles and buttons (toggle switches take effect immediately vs checkboxes require submission · split buttons: default action + arrow menu with <10–12 items · cancel vs close X icon ambiguity to prevent data loss) ④ Copy and login (link label 4S: Specific/Sincere/Substantial/Succinct · passwordless OTP/Passkey: 76% store credentials on phones/83% use biometrics) ⑤ Multi-step flows and tools (8 wizard guidelines + limitations · drag-and-drop magnetic snap with 100ms animation · progressive disclosure for accelerators · three calculator quiz types: conversion/prediction/recommendation + 12 guidelines + user expectation of 2–7 inputs). Audience = product managers; terms explained inline; decision-oriented guidance on when to use/how to choose/counterexamples.