魔法墨水

9 min · 8 分钟English · 中文

Magic Ink魔法墨水

Information software is graphic design, not interaction design

信息软件的本质是平面信息设计,而非交互设计

An interactive English–Chinese guide to Bret Victor’s Magic Ink: why every click is a cost, and how context inference, graphic density, and direct manipulation replace interrogation UIs.

以中英双语和交互对照实验解读 Bret Victor 的《Magic Ink》:为什么每次点击都是成本,以及语境推断、图形密度与直接操纵如何取代审讯式界面。

Will the ride home be dry?骑车回家,会遇上雨吗?
Interrogation vs. Magic InkOne evening forecast aligns temperature and rain. The dry window lasts from 19:00 to 20:30.
Temp. °C Rain0 clicks · 1 glance

Information software is not a machine

信息软件不是一台机器

You open an app because you want to know one simple thing: whether to pack a rain jacket before biking to dinner, which movie fits your evening, or which book is worth your weekend. You do not wake up hoping to operate a control panel. You want to learn something, decide, and get back to your life.

你打开一个应用,通常只是想知道一件简单的小事:晚上骑车去吃饭要不要带雨衣,今晚哪场电影接得上日程,或者周末该读哪本书。你醒来并不是为了操作一块仪表盘。你是为了弄明白一件事,做出决定,然后回到生活里去。

In his landmark 2006 essay Magic Ink: Information Software and the Graphical Interface, Bret Victor argues that most of the software we use every day is designed around the wrong metaphor. We build tools for reading and learning as if they were machines for operating and editing.

在 2006 年那篇影响深远的长文 Magic Ink:信息软件与图形界面 中,Bret Victor 指出,我们每天使用的大多数软件,从一开始就套错了隐喻。我们把用来阅读与学习的工具,设计成了用来操纵与编辑的机器。

Victor begins by dividing all software into three categories according to what the human is actually trying to do:

Victor 首先按人的真实目的,把所有软件分成三类:

  • Manipulation software helps you create or change an artifact — writing a document, editing a photo, drafting a circuit. Here, direct control and rich tools are the whole point.
  • Communication software connects you with other people — email, chat, calls. The interface is a conduit between minds.
  • Information software helps you learn about the world and make decisions — weather, maps, schedules, news, shopping, reference, search. Most of the web lives here.
  • 操纵型软件(Manipulation)帮助你创作或修改某件作品——写文档、修照片、画电路。在这里,精细的控制和工具本身就是目的。
  • 沟通型软件(Communication)把你和他人连接起来——邮件、聊天、通话。界面是人与人之间的通道。
  • 信息型软件(Information)帮助你了解世界并做出决定——天气、地图、时刻表、新闻、购物、检索与百科。互联网的大部分内容都属于这一类。

The graphical user interface was invented for manipulation software: buttons, menus, forms, dialogs, toolbars. When the web arrived, designers simply copied that vocabulary onto information software. Instead of showing people a clear picture, we handed them a cockpit and asked them to interrogate a database one click at a time.

图形用户界面原本是为操纵型软件发明的:按钮、菜单、表单、对话框、工具栏。等到万维网兴起,设计师几乎不假思索地把整套控件搬进了信息软件。我们没有给人看一张清楚的图,而是塞给他一座驾驶舱,让他一次次点击,像审讯数据库一样套出答案。

Information software design is primarily graphic design, not interaction design. Show people the answer instead of making them ask for it.
信息软件设计本质上是平面信息设计,而不是交互设计。把答案直接画出来,别逼人一步步去问。

Three kinds of software, one misplaced metaphor三种软件,一个错位的隐喻

Learning 阅读理解Manipulating 动手操作

Checking the evening train timetable. 查看今晚的列车时刻表。

Design fit 合宜设计
Align departures and arrivals so the eye can compare trips. 对齐发车与到站时间,一眼就能比较班次。
Common mistake 常见误区
Each departure opens on a separate detail page. 每趟车都要点进单独的详情页。

Historical mismatch 历史错位

ManipulationInformation
Buttons, menus, forms and dialogs cross into the timetable. 按钮、菜单、表单和弹窗,一并被搬进了时刻表。

Magic Ink path 魔法墨水的路径

  1. Graphic design 图形设计
  2. Context inference 推断语境
  3. Earned direct manipulation 必要时直接操作
The bars show the emphasis of each example. In communication software, reading and composing serve the exchange.条带表示各例中活动的侧重。在沟通型软件里,阅读与撰写都服务于人与人的交流。

Interaction is a last resort

交互是最后的退路

Victor’s sharpest claim is easy to misread until you feel it yourself: in information software, interaction is a last resort. Every click, dropdown, tab, and form field is not a feature — it is a toll. It is a concession the designer makes when the display failed to put the right information in front of the reader’s eyes.

Victor 最锋利的论断,在亲手体验之前很容易被误解:在信息软件里,交互是最后的退路。每一次点击、下拉框、标签页和表单输入,都不是产品的加分项,而是用户付出的通行费——是界面没能把关键信息直接摆到眼前时,设计师不得不做出的妥协。

Think about why the finest artifacts of print graphic design — transit maps, train timetables, financial broadsheets, Edward Tufte’s charts — are still astonishingly fast to read. Paper has zero interactivity. A print designer cannot hide tomorrow’s rain behind a tab or tuck half the schedule onto page two. Forced to fit everything onto one motionless sheet, graphic designers learned to layer context, alignment, and comparison so the human eye could extract an answer in seconds.

想想为什么最出色的纸质平面设计——地铁图、列车时刻表、财经报纸、Edward Tufte 推崇的统计图表——至今依然读得飞快。因为纸张的交互性为零。纸面设计师没法把明天的降雨藏进标签页,也没法把一半时刻表塞进第二页。正是因为必须把一切放进同一张静止的画面,平面设计师磨炼出了分层、对齐与并置的技艺,让人的眼睛几秒钟就能扫出答案。

Software removed that constraint, and designers got lazy. Whenever a layout felt crowded, the reflex was to add a tab, a modal, a filter drawer, or a search box. We replaced clear graphics with interactive crutches.

软件撤掉了纸面的限制,设计师反而变懒了。只要版面显得有点挤,第一反应就是加标签页、加弹窗、加筛选抽屉、加搜索框。我们用交互的拐杖,替代了本该画清楚的图形。

Victor offers a better mental model: magic ink. Imagine ink on paper that can dynamically redraw itself for your specific situation right now. You would never design a printed weather chart with a “Next 3 Hours” button. You would design the single richest, cleanest graphic that answers the reader’s question at a glance — and add interaction only where a static picture genuinely runs out of room.

Victor 给出了一个更好的心智模型:魔法墨水(Magic Ink)。想象纸上的墨水能够根据你此刻的具体处境,自动重绘成最贴合的画面。你绝不会在纸质天气图上印一个“查看后三小时”的按钮;你会画出那张信息最饱满、结构最清澈的图,让人一眼看懂——只有当静态画面真的装不下时,才谨慎地引入交互。

Try the comparison below on a real question. Feel the difference in your wrist and your head between interrogating a tabbed weather app and reading one sheet of magic ink.

试试下面这个真实任务的对比。在手腕和大脑里亲自感受一下:在标签页迷宫里拼凑天气,与直接读一张魔法墨水图,究竟差了多少。

The weather test: Click maze vs. magic ink天气测试:点击迷宫 vs. 魔法墨水

Task: You’re biking to dinner at 7 PM and riding home at 10 PM. Do you need a rain jacket, and when is the warmest dry stretch?

任务:你晚上 7 点骑车去吃晚饭、10 点骑车回家。要不要带雨衣?最暖和又干燥的时段是几点?

Local weather · 本地天气
2 PM · 当前实况21°C

Partly Cloudy · 局部多云

Watch these isolated readings fly onto one continuous evening curve · 看零散读数飞上同一条傍晚曲线
Would you pack a jacket? · 你会带雨衣吗?

Illustrative forecast. All times are local.模拟天气预报,时间均为当地时间。

Infer from context

从语境中推断

If interaction is a cost, how do you actually eliminate it without drowning the screen? Victor’s first lever is context-sensitive inference: stop asking the user to type in facts the software already knows or can reasonably guess.

既然交互是成本,怎样才能在不把屏幕塞爆的前提下省掉它?Victor 的第一根杠杆是基于语境的推断:别再让用户手动输入软件本来就知道、或者完全可以合理猜到的信息。

Most web interfaces treat every visit like an interrogation at customs. Where are you? Where are you going? What time is it? Do you want hourly or daily? Forms force you to translate your lived situation into database fields before the screen will tell you anything useful.

大多数网页把每一次访问都当成海关盘查:你在哪儿?要去哪儿?现在几点?要看逐小时还是十天预报? 表单逼着你先把自己鲜活的处境翻译成数据库字段,屏幕才肯吐出一点有用的东西。

Magic ink starts from context — your location, the clock on the wall, what you were just doing, the structure of the task — and renders the best answer immediately. When an assumption might be wrong, you do not block the screen with a search form; you show the assumed context softly in the margin so a single tap can adjust it while the graphic stays live.

魔法墨水则从语境出发——你的位置、墙上的时钟、你刚看过的东西、任务本身的结构——立刻画出最合理的答案。如果某个猜测可能不准,也绝不用搜索表单拦住整块屏幕,而是把当前假设轻巧地写在边缘:点一下就能微调,画面始终保持鲜活。

Don’t ask the user to describe their situation when the software can look around and start with a good guess.
软件只要环顾一下就能猜个八九不离十,就别再逼用户自己填表描述处境。

Compare the two transit boards below. One makes you fill out a form and click through route details; the other simply looks at where you are and what time it is, and draws the race between your walk and the train.

对比下面两块出行看板。左边逼你填表单、点开一条条路线详情;右边直接读出你的位置与时间,把你步行与列车的赛跑画在了同一根时间轴上。

Context inference: Form interrogation vs. living departure board语境推断:表单盘查 vs. 活的出发看板

Task: It’s 5:42 PM near Zhongshan Park and you’re heading to Hongqiao Terminal 2. Can you catch the 5:54 train, or should you grab coffee and take the 6:06?

任务:现在 17:42,你在中山公园附近,要赶去虹桥 2 号航站楼。赶得上 17:54 那班车吗,还是干脆买杯咖啡坐 18:06?

Transit enquiry · 路线查询

Simulated timetable. Doors close two minutes before departure. / 模拟时刻表,发车前两分钟关门。

Show the whole space

展开整个决策空间

Victor’s second lever is graphic density: show the whole decision space at once so the human visual system can do the filtering.

Victor 的第二根杠杆是图形密度:把整个决策空间一次铺开,把筛选的工作交给人的视觉系统。

Your eyes are staggeringly fast parallel processors. Glance at a well-drawn timeline or map and you simultaneously compare distances, durations, gaps, and clusters without conscious effort. Clicking across separate screens destroys that gift. Every time you open a detail page and hit Back, you force comparisons through tiny working memory — trying to remember whether the first movie started at 7:45 or 8:05 while staring at the second one.

人眼是惊人的并行处理器。扫过一张画得好的时间轴或地图,你几乎不费意识就能同时比较距离、时长、空档与疏密。而跨页面点击会彻底毁掉这种天赋。每当你点进详情页再按返回,所有的比较都被迫挤进狭小的工作记忆——盯着第二部电影时,还得硬记第一部到底是 7:45 还是 8:05 开场。

One of Victor’s most memorable redesigns in Magic Ink is a local movie listing. Typical sites organize by movie or by theater, locking each showtime inside a separate page. Victor flattens the entire evening onto one unified visual schedule: every film as a bar of its true length, anchored to a shared clock, paired with how long it takes to get there and whether it is worth seeing.

《Magic Ink》里最令人难忘的改造之一,就是本地电影排片页。常见网站要么按影片分类、要么按影院分类,把每一场次关进各自独立的页面。Victor 把整晚的排片压平到同一张时间图里:每部片子按真实片长画成一条横杠,对齐同一根时钟轴,顺手标出路程耗时与评分。

When all the options share one visual coordinate system, you stop navigating and start seeing. Tradeoffs that were invisible across six tabs leap off the page in a single scan.

当所有选项共享同一套视觉坐标系,你就不再是在“导航”,而是在看见。原本藏在六个标签页背后、怎么也比不清的取舍,扫一眼就浮出水面。

Try finding a movie that fits your evening in both versions below. Notice how quickly the paginated cards force you to juggle facts in your head — and how effortlessly the unified timeline settles the choice.

在下面两个版本里试着挑一场契合今晚日程的电影。留意分页卡片多快就开始折磨你的短期记忆,而统一时间轴又是如何一眼定音的。

Showing the whole space: Movie cards vs. evening timeline展开整个空间:分页电影卡 vs. 晚间时间轴

Task: Find a movie rated 8.0+ starting after 7:30 PM, within a 15-minute trip, that finishes before 10:30 PM.

任务:找一部评分 8.0 以上、19:30 后开场、路程不超过 15 分钟、且 22:30 前散场的电影。

Page 1 of 2 · 第 1 / 2 页
Watch these cards align with all 14 showings on one clock · 看这些卡片直接对齐到同一晚的 14 个场次

When interaction is earned

当交互物有所值

Victor is not arguing that information software should never be interactive. Sometimes you genuinely do not know what you want until you see what is available — choosing a book, hunting for an apartment, or weighing flights where price, duration, and timing trade off against one another.

Victor 并不是说信息软件永远不该交互。有些时候,你在看到有哪些选项之前,本来就不清楚自己真正要什么——挑一本书、租一套房、选一趟在价格与转机之间权衡的航班,你的偏好是在浏览的过程中逐渐成形的。

What Victor attacks is turn-based interaction: the stiff, bureaucratic loop of fill out form → click Apply → wait for a new list → realize you filtered too hard → hit Back → tweak a checkbox → submit again. Turn-based filters blindfold you between moves. Worse, when your criteria leave zero results, they hand you an empty screen and hide the wonderful option sitting two steps outside your cutoff.

Victor 反对的是回合制交互:那种僵硬官僚的循环——填表 → 点“应用筛选” → 等列表刷新 → 发现筛得太死 → 返回修改 → 再次提交。回合制筛选在每一步之间蒙住你的眼睛;更糟的是,一旦条件太苛刻,它只丢给你一个“0 条结果”的空白页,把离阈值只差一步的好选项藏得严严实实。

When interaction is truly earned, it should feel like direct, continuous manipulation of the picture itself. Move a slider and watch the graphic respond on every pixel of motion. Give the controls scent — tiny embedded histograms showing where items actually live — so you never drag blindly into empty space.

当交互真的不可或缺时,它应当像直接、连续地操纵画面本身。拖动滑块,图形随着每一个像素的移动实时重绘。给控件加上气味(scent)——在滑轨里嵌入小小的分布直方图,让你一眼看见数据聚在哪里,不必盲目拖进荒原。

And most importantly: dim non-matching options instead of erasing them. When a great book sits eighteen pages past your limit, keeping it softly visible in the background teaches you something a blank list never could.

最关键的是:把不满足条件的选项变暗,而不是删掉。当一本极好的书只是比你的页数上限多出十八页时,让它在背景里隐约可见,能教会你空白列表永远教不会的事。

Explore the bookshop both ways below. Try tightening your constraints until the traditional filter hits a dead end — then drag the live sliders and watch how the near-misses stay legible.

用下面两种方式逛一逛书单。试试把传统筛选调严,直到撞上“无结果”的死胡同——再拖动魔法墨水的实时滑块,看看那些近在咫尺的遗珠如何留在视线里。

Earned interaction: Turn-based filter vs. scented live landscape值得的交互:回合制筛选 vs. 实时气味图景

Task: Find a short (< 250 pages), highly rated design or computing book that leans accessible rather than academic — and notice what sits just outside your page limit.

任务:找一本篇幅较短(250 页以内)、口碑极佳、偏通俗而非学院派的设计或计算书——再看看略微超出页数上限处藏着什么好书。

Reading tone · 阅读难度

4 books matched · 符合条件的书籍 4 本

  • Don’t Make Me Think点石成金
    Steve Krug · 216 p / 页 · ★ 4.6
  • Thinking in Systems系统之美
    Donella Meadows · 240 p / 页 · ★ 4.8
  • Understanding Comics理解漫画
    Scott McCloud · 215 p / 页 · ★ 4.9
  • Mindstorms头脑风暴
    Seymour Papert · 230 p / 页 · ★ 4.7
Watch these books fly onto a 2D landscape — with near-misses kept softly visible · 看这些书飞上二维图景,并把遗珠留在视线内

What options look like in magic ink

用魔法墨水看懂期权

If you have never traded an option before, start with one simple idea: an option is a dated side-contract on a stock. For an upfront ticket price called the premium, you buy the right (never the obligation) to trade 100 shares at a locked-in strike price on or before an expiration date. Because your downside is capped at what you paid for the ticket, options come in two mirror flavors and a natural midpoint:

如果你从未交易过期权,可以先记住一个最简单的画面:期权是一张附带期限的约定门票。你先付一笔门票费,叫做权利金,就能换取在到期日当天或之前,按一个事先锁定的行权价买卖股票的权利(而不是义务)。因为最大亏损就止于门票费,期权自然分成一对镜像,并在现价附近有一个天然中轴:

  • Call (▲) — the right to buy the stock at the strike price. You want the stock to rise far enough above the strike to pay back the ticket (strike + premium).
  • Put (▼) — the right to sell the stock at the strike price. You want the stock to fall far enough below the strike to pay back the ticket (strike − premium).
  • ~0.5 Delta (Δ) — a contract whose strike sits right near today’s stock price, where the market treats finishing in the money as roughly a 50/50 coin flip.
  • 看涨期权 Call(▲)——按行权价买入股票的权利。到期股价必须涨过行权价并赚回门票费(行权价 + 权利金),你才真正盈利。
  • 看跌期权 Put(▼)——按行权价卖出股票的权利。到期股价必须跌破行权价并赚回门票费(行权价 − 权利金),你才真正盈利。
  • 约 0.5 Delta(Δ)——行权价紧贴今日股价的合约。市场眼里,它到期落入价内与价外的概率大致各占一半。

Financial software turns this straightforward idea into one of the most intimidating cockpits on the web. Open a brokerage options screen and you face an option chain — fifteen strikes, bids, asks, open interest, implied volatility, and Greek letters — paired with an isolated P&L diagram for a single contract. To look across time, you click one expiration tab after another, trying to hold each spread in your head.

传统金融软件把这么朴素的逻辑,做成了一块令人生畏的驾驶舱仪表盘。打开任何券商期权界面,迎面而来就是密密麻麻的期权链——十五档行权价、买卖报价、未平仓量、隐含波动率和一串希腊字母——再配上一张孤零零的单合约到期盈亏图。想看清不同时间跨度?你只好一个个点开到期日标签页,靠大脑硬记刚看过的数字。

None of those tabs shows what you actually came to understand. A single-contract payoff hockey stick tells you what happens if you buy one strike on one Friday and wait until expiration. Detached from the stock’s own price history and stripped of neighboring dates, options look like opaque lottery tickets rather than a coherent market.

这些标签页偏偏没有画出你真正想弄明白的那件事。单合约的“曲棍球棒”盈亏图,只告诉你押注某一个到期日、某一个行权价会怎样。既看不见股票自身的价格走势,也看不见相邻到期日之间的关系,期权于是显得像一堆晦涩的彩票,而不是一个浑然一体的市场。

Now try painting the whole options market onto one sheet of magic ink. Keep the stock’s historical price on the left side of the chart, and place future expiration dates along the same timeline on the right. Then lift just that highlighted ~0.5 delta row from each expiry tab — one call (▲) and one put (▼) per expiration date.

现在试着把整片期权市场画进同一张魔法墨水图里。左边留给股票的历史价格曲线,右边沿着同一根时间轴铺开未来的各个到期日。然后,从每一个到期日标签页里,只抽出那一行高亮的 ~0.5 Delta 合约对——每个到期日各取一张看涨期权(▲)和一张看跌期权(▼)。

If you first plot those contracts at their strike price, nothing interesting appears: every green call triangle and red put triangle lines up flat against today’s price. Now apply the basic logic of an option. To actually profit by expiration, a call buyer has to clear the strike plus the premium paid (strike + call), while a put buyer has to fall below the strike minus the premium (strike − put). Lift every call up to its breakeven and drop every put down to its breakeven — and watch what happens.

如果先把这些合约画在各自的行权价上,画面毫无波澜:所有绿色看涨三角和红色看跌三角只是在现价水平排成一条直线。接着,把期权最朴素的逻辑放上去:看涨期权买家要真正回本,到期股价必须越过行权价 + 权利金;看跌期权买家要真正回本,股价必须跌破行权价 − 权利金。把每个看涨三角抬高到它的盈亏平衡点,把每个看跌三角降到它的盈亏平衡点——看看画面里长出了什么。

The moment every contract moves to its breakeven, a curved cone blooms out of the price chart. Options stop looking like isolated bets and reveal what they truly are: a prediction market for volatility.
当每一张合约都移到自己的盈亏平衡点,一只优雅展开的圆锥便从股价末端生长出来。期权不再像一张张孤立的赌票,而瞬间显露出它的真实面目:一个为波动率定价的预测市场。

Because uncertainty accumulates roughly with the square root of time, those breakevens flare outward rapidly over the next few weeks and curve more gently into later years — forming the exact envelope of where the crowd thinks the stock might plausibly travel. Experience the three tabs below from left to right: start inside the 570-contract traditional chain and single payoff chart, step to the middle tab to isolate the glowing $125 (~0.5Δ) row across expiries, and flip into magic ink to watch those strikes bloom into a living volatility cone.

因为未来的不确定性大致随时间的平方根累积,这些盈亏平衡点在近期迅速张开,向远期逐渐趋缓,恰好勾勒出市场群体眼中股价未来可能的边界。从左到右依次点开下面三个标签页:先在包含 570 张合约的传统期权链与单合约盈亏图里感受迷雾,接着切到中间页横向抽取出每个到期日高亮的 $125(~0.5Δ)平值行,最后翻进魔法墨水视图——亲眼看它们顺着时间线绽放成一只活的波动率圆锥。

Seeing options: Chain & single P&L vs. breakeven cone看懂期权:期权链与单合约盈亏 vs. 盈亏平衡圆锥

Task: Judge how far the market thinks VCTR ($125) could swing over the next two years — watching the same $125 (~0.5Δ) contracts transform from 570 chain cells into a breakeven volatility cone.

任务:判断市场预计 VCTR(现价 $125)未来两年的潜在波动区间——看同一组 $125(~0.5Δ)合约如何从 570 格传统期权链中脱离出来,一步步变形为盈亏平衡圆锥。

New to options? 4 essentials in 20 seconds · 期权零基础速览

Call ▲ · 看涨(买入权)

Right to buy at the strike. Profits only above Strike + Premium.

按行权价买入的权利;到期涨过「行权价 + 权利金」才盈利。

Put ▼ · 看跌(卖出权)

Right to sell at the strike. Profits only below Strike − Premium.

按行权价卖出的权利;到期跌破「行权价 − 权利金」才盈利。

Strike & Premium · 行权价与权利金

Strike (K) is the contract’s locked price; Premium is the upfront ticket cost.

行权价是合约锁定的交易价,权利金是买下这张合约的门票费。

~0.5 Delta (Δ) · 平值约 0.5Δ

The strike sitting right near today’s $125 price — roughly a 50/50 coin flip at expiry.

紧贴今日 $125 现价的行权价,代表到期落在价内价外约各半的中轴。

VCTR Option Chain & Single-Contract Payoff · 期权链与单合约盈亏15 strikes × 2 sides × 19 expiries = 570 contracts · Mar 17, 2028 (382d)
Click any of the 30 contracts on this tab · 点击任一合约Highlighted: ~0.5Δ ATM pair · 仅平值行将进入下一阶段
OIΘ/dΔIVCall · 看涨Strike · 行权价Put · 看跌IVΔΘ/dOI
120−$0.02+0.8247%$9047%-0.17−$0.02120
160−$0.02+0.7947%$9547%-0.20−$0.02260
570−$0.02+0.7647%$10047%-0.23−$0.021,090
810−$0.03+0.7346%$10546%-0.26−$0.021,540
2,010−$0.03+0.7046%$11046%-0.29−$0.023,180
2,730−$0.03+0.6746%$11546%-0.32−$0.023,610
4,230−$0.03+0.6345%$12045%-0.36−$0.035,050
9,070−$0.03+0.6045%$125~0.5Δ ATM45%-0.39−$0.038,260
5,670−$0.03+0.5645%$13045%-0.42−$0.033,930
3,570−$0.03+0.5345%$13545%-0.46−$0.032,610
3,220−$0.03+0.5044%$14044%-0.49−$0.021,970
1,600−$0.03+0.4744%$14544%-0.52−$0.02810
1,160−$0.03+0.4444%$15044%-0.55−$0.02500
280−$0.03+0.4144%$15544%-0.58−$0.02150
120−$0.03+0.3843%$16043%-0.61−$0.02120

Mar 28 · $125 Call 看涨

Spot $125 · IV 45%
Long call at $125 strike costing $23.40. Breakeven at $148.40.+$23.4BE $148.4K $125$0−$23+$35$80$100$120$140$160$180
Strike · 行权价
$125
Premium · 权利金
$23.40
Breakeven · 盈亏平衡
$148.40
Delta (Δ) · 方向弹性
+0.60
Theta (Θ/day) · 日损耗
−$0.03
Vega (+1% IV) · 波动弹性
+$0.49

To compare expiries in this view, you flip through 570 contracts one tab at a time. Next, zoom out so this table shrinks into its slot alongside five neighboring expiries and highlights only the $125 (~0.5Δ) row:想比较不同到期日,传统视图只能在一张 570 合约的大表里反复切标签。下一步,将这张表缩小放进时间轴,与相邻到期日并排铺开,并只点亮 $125(~0.5Δ)平值行:

Illustrative Black-Scholes pricing on a simulated stock (VCTR, spot $125) with modest carry (r = 2.5%, q = 1.0%) and a gentle post-selloff volatility term structure.基于模拟标的 VCTR(现价 $125)的 Black-Scholes 示例定价,含温和持有成本(r = 2.5%, q = 1.0%)与急跌后的期限波动率结构。

Start with paper, earn every widget

从纸面开始,让每个控件自证其价

Read Magic Ink twenty years after it was written and its lesson feels more urgent than ever. We spent a decade wrapping information in hamburger menus, paginated cards, and multi-step wizards. Today, we are tempted to wrap everything in a chat box — arguably the purest turn-based interrogation interface ever invented. Asking an assistant three back-and-forth questions to discover a schedule or compare tradeoffs is just a friendlier form of clicking Submit.

二十年后再读《Magic Ink》,它的提醒反而比当年更迫切。我们花了一整个十年,把信息包进汉堡菜单、分页卡片和多步向导里;今天,我们又忍不住想把一切都包进对话框——那几乎是人类发明过最纯粹的回合制审讯界面。为了查一个时刻表或比较几种取舍,去跟助手你一言我一语聊上三轮,不过是把点“提交”换成了打字。

Yet the same technology that makes chat easy also makes Victor’s dream finally practical: software that can synthesize context on the fly and paint bespoke, high-density visual artifacts for the exact question in front of you.

但让对话变得轻而易举的同一代技术,也第一次让 Victor 的梦想触手可及:软件可以即时综合语境,为你眼前这个具体问题,现画出一张高密度、可一眼读懂的专属图形。

Before adding another button, filter drawer, or prompt loop to an information product, Victor’s essay leaves us with three quiet tests:

下次准备给信息类产品再加一个按钮、筛选抽屉或对话轮次之前,Victor 这篇文章留下了三个安静的检验:

  1. What question is the reader actually trying to answer? Design the ideal motionless graphic for that question first, as if it had to be printed on paper.
  2. What is the interface asking that it already knows? Replace form fields with gentle, inspectable defaults inferred from time, place, and task.
  3. What is hidden behind clicks that the eye could compare at once? Put competing options into one shared visual frame — and when the reader truly needs to steer, let them sculpt the picture continuously instead of submitting forms.
  1. 读者此刻真正想回答什么问题? 先假设它只能印在纸上,为这个问题画出最理想的静止图形。
  2. 界面在追问哪些它本来就知道的事? 用基于时间、地点与任务的柔和默认推断,取代拦路的表单字段。
  3. 有哪些被藏在点击背后的选项,其实可以让眼睛一眼并排比较? 把竞争选项放进同一个视觉坐标系——只有当读者确实需要掌舵时,才让他连续地雕塑画面,而不是反复交表。
“Design is the craft of showing people what they need to see.” — Inspired by Bret Victor’s Magic Ink (2006).
“设计的技艺,是把人需要看见的东西呈现出来。” —— 写于重读 Bret Victor 《Magic Ink》(2006)之后。

Field guide: Interrogation UI vs. Magic Ink速查表:审讯式界面与魔法墨水

Starting point 设计起点

Interrogation UI 审讯式界面

Start with screens, menus, forms, and flows borrowed from editing tools.

从编辑软件借来页面、菜单、表单与跳转流程。

Magic Ink 魔法墨水

Design the ideal static graphic first, as if it had to work on paper.

先假设只能印在纸上,画出最理想的静止信息图。

Forces clarity and density before adding a single widget. 在添加任何一个控件之前,先逼出清澈与密度。

Role of interaction 交互的角色

Interrogation UI 审讯式界面

Hide details behind tabs, accordions, modals, and pagination.

把细节藏进标签页、折叠面板、弹窗与分页里。

Magic Ink 魔法墨水

Treat every click as friction; let the human eye scan and compare.

把每次点击都视为摩擦,把扫描与比较交给眼睛。

Vision compares in parallel; clicking bottlenecks working memory. 视觉可以并行比较,跨页点击会挤爆工作记忆。

User context 用户语境

Interrogation UI 审讯式界面

Greet visitors with an empty search box or mandatory query form.

一打开就用空白搜索框或必填查询表单拦住去路。

Magic Ink 魔法墨水

Infer place, time, and task immediately; keep assumptions softly editable.

立刻推断地点、时间与任务,把假设留在边缘方便轻点微调。

Eliminates interrogation before the user has even seen the data. 省掉用户还没看到数据就得先填表的盘查感。

Comparing options 比较选项

Interrogation UI 审讯式界面

Separate detail pages that force clicking in and hitting Back.

分散的详情页,逼着人反复点进又点返回。

Magic Ink 魔法墨水

Align all candidates on one shared axis, map, or visual grid.

把所有候选放进同一根坐标轴、地图或视觉网格里。

Tradeoffs and outliers jump out in a single glance. 取舍与异常值扫一眼就浮出水面。

Filtering & steering 筛选与掌舵

Interrogation UI 审讯式界面

Turn-based checkboxes + “Apply” button that erase non-matches into empty states.

勾选复选框 + 点“应用筛选”,把不匹配项彻底删成空白页。

Magic Ink 魔法墨水

Continuous, scented sliders that update live and dim near-misses in place.

带分布气味的连续滑块,实时重绘并原地变暗遗珠而非删除。

Reveals cliffs, clusters, and great options sitting just past a cutoff. 让人看见断崖、聚类,以及离阈值只差一步的好选择。