中文

1981 年的变化不一定显眼，但它很“决定命运”。如果说 1980 还在回答“计算能不能进入组织”，那么 1981 开始回答一个更具体、更难的问题：它能不能被真正用起来，并且被更多人用起来？

技术史常被写成机器的故事。但对多数组织而言，机器只是开始。真正的难点在后面：数据怎么进来、错误怎么发现、流程怎么衔接、结果怎么被信任。

1. 全球切片：从“能买到”走向“能长期用”

在全球视角里，个人计算的轮廓继续变得清晰：硬件变得更像商品，软件开始更像产业，兼容与标准的压力越来越真实。很多看似细碎的改进（更好的工具、更可预期的使用方式、更容易学习的操作）会把计算从少数人的玩具，推向更广的日常。

但真正决定扩散速度的，往往不是某个性能指标，而是：

• 工具链是否完整：写作、表格、数据处理、打印、存储与备份有没有“默认路径”。
• 失败是否可控：崩溃、丢失、格式不兼容，会不会让人不敢依赖。
• 迁移成本是否上升：一旦开始积累数据与文件，换系统就会变得昂贵。

2. 国内切片：从“调用能力”走向“训练技能”

在国内，计算仍然更多存在于机构、科研、工程与教育体系里。它的扩散方式更像“技能的扩散”，而不是“商品的扩散”。

这会自然地把注意力推向机房与制度：

• 谁可以上机，按什么规则上机
• 上机前要准备什么（纸面推演、算法、流程）
• 出错后如何定位（是数据错、逻辑错、操作错，还是系统限制）

当上机成为一套可训练、可重复的流程，计算就不再是“少数人的神秘能力”，而开始成为一种可以被组织吸收的技能。

3. 关键转折：录入、流程、纠错，决定“可用性”

1981 这一章的核心词是：可用性（usable）。它不是界面漂亮不漂亮，而是系统能否在真实世界里被反复使用。

3.1 录入：一切系统的瓶颈

系统要运行，数据必须进入系统。录入谁来做？怎么校验？录错了怎么回滚？

很多项目失败并不是“算不出来”，而是“录不进去”“录进去也不可信”。一旦录入成本过高，组织就会回到纸与人。

3.2 流程：计算必须嵌入工作，而不是悬在工作之上

如果计算结果不能自然进入下一步流程（审批、生产、采购、统计），它就只能停留在演示与试点。

当流程开始围绕数据重新排列，计算才真正进入组织。

3.3 纠错：让人敢于依赖

错误是必然的：数据会错，人会错，系统也会错。关键在于：错误是否可追踪、可修正、可解释。

可用性的成熟，意味着组织开始建立一套“错误管理”的文化：

• 错误被记录，而不是被隐藏
• 责任被界定，而不是被推诿
• 结果可复查，而不是凭权威

4. 国内+全球对照：同一条路，不同的起点

• 全球更像“市场推动的下沉”：工具变便宜、变好用、生态更强。
• 国内更像“组织推动的扩散”：通过教育与制度，把计算作为技能训练出来。

两条路会在后面相遇：当商品下沉与技能扩散交汇时，计算才会真正进入大众生活。

5. 结尾：从“能算”到“能用”，决定下一阶段

1981 的意义，是把叙事从“机器出现”转向“系统可用”。从这一年开始，计算不再只是技术能力，而越来越像一种组织能力：能否让更多人参与，能否让错误可控，能否让流程可持续。

下一章（第三章，1982）我会继续沿着这条线写：当可用性成为目标，工具链、标准化与协作方式会怎样开始塑形。

---

English

In 1981, change may not look dramatic, but it becomes decisive. If 1980 still asks whether computing can enter an organization at all, 1981 begins to ask a harder, more practical question: can it be used—repeatedly, reliably, and by more people?

Technology history is often told as a story of machines. For most organizations, the machine is only the beginning. The real difficulty comes later: how data gets in, how errors are discovered, how workflows connect, and how results become trustworthy.

1. A global slice: from “buyable” to “livable”

In the global view, personal computing continues to acquire a stable outline. Hardware behaves more like a product category; software begins to behave more like an industry; compatibility and standards start to feel like pressure rather than theory.

What decides the speed of adoption is rarely a single benchmark. It is usually:

• whether the toolchain is complete (writing, tables, printing, storage, backup)
• whether failure is survivable (crashes, loss, incompatible formats)
• whether migration costs rise as data accumulates

2. A China slice: from “requested capability” to “trained skill”

In China, computing still lives primarily inside institutions, research, engineering, and education. Its diffusion resembles the diffusion of skill rather than the diffusion of commodities.

That pulls attention toward the computer room and its disciplines:

• who gets machine time, and under what rules
• what preparation happens before the session
• what counts as an error, and how it is traced

Once “going on the machine” becomes trainable, repeatable practice, computing stops being a mysterious power of a few. It becomes something an organization can absorb.

3. The turning point: input, workflow, error handling

The keyword of 1981 is usability—not aesthetic polish, but the ability to survive reality.

3.1 Input: the bottleneck of every system

No system runs without data. Who enters it? How is it verified? How do you correct it without breaking everything downstream?

Many projects fail not because they cannot compute, but because they cannot ingest trustworthy data at sustainable cost.

3.2 Workflow: computing must sit inside work, not above it

If results cannot flow naturally into the next step—approval, production, procurement, reporting—computing remains a demo.

Computing enters an organization only when workflow begins to reorganize around data.

3.3 Errors: the moment people dare to rely

Errors are inevitable: data errs, people err, systems err. The question is whether errors are traceable, correctable, and explainable.

Usability matures when an organization builds a culture of error management:

• errors are recorded rather than hidden
• responsibility is defined rather than deferred
• results are inspectable rather than purely authoritative

4. Contrast: the same road, different starting points

• Globally, the descent is market-driven: tools become cheaper, more usable, and ecosystem-rich.
• In China, the spread is organization-driven: computing is trained as a skill via education and rules.

The two paths later converge. When commodity access meets skill diffusion, computing finally becomes ordinary.

5. Closing: from calculable to usable

The meaning of 1981 is a shift in narrative: from “a machine exists” to “a system can be used.” Computing begins to look less like a technical capability and more like an organizational capability—scalable participation, controllable errors, sustainable workflow.

Next (Chapter 3, 1982), I will follow the same line: once usability becomes the goal, toolchains, standardization, and collaboration begin to take shape.