惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

博客园_首页
C
Cybersecurity and Infrastructure Security Agency CISA
C
Cyber Attacks, Cyber Crime and Cyber Security
Project Zero
Project Zero
P
Proofpoint News Feed
D
Darknet – Hacking Tools, Hacker News & Cyber Security
C
Cisco Blogs
V
Vulnerabilities – Threatpost
G
GRAHAM CLULEY
N
News | PayPal Newsroom
NISL@THU
NISL@THU
雷峰网
雷峰网
J
Java Code Geeks
Latest news
Latest news
aimingoo的专栏
aimingoo的专栏
Microsoft Azure Blog
Microsoft Azure Blog
OSCHINA 社区最新新闻
OSCHINA 社区最新新闻
Cisco Talos Blog
Cisco Talos Blog
Hacker News: Ask HN
Hacker News: Ask HN
AWS News Blog
AWS News Blog
Application and Cybersecurity Blog
Application and Cybersecurity Blog
T
The Exploit Database - CXSecurity.com
P
Privacy International News Feed
C
CXSECURITY Database RSS Feed - CXSecurity.com
Vercel News
Vercel News
Spread Privacy
Spread Privacy
V2EX - 技术
V2EX - 技术
S
Schneier on Security
K
Kaspersky official blog
Recent Announcements
Recent Announcements
T
Threat Research - Cisco Blogs
B
Blog RSS Feed
S
SegmentFault 最新的问题
Security Archives - TechRepublic
Security Archives - TechRepublic
Stack Overflow Blog
Stack Overflow Blog
Hugging Face - Blog
Hugging Face - Blog
Apple Machine Learning Research
Apple Machine Learning Research
Threat Intelligence Blog | Flashpoint
Threat Intelligence Blog | Flashpoint
W
WeLiveSecurity
PCI Perspectives
PCI Perspectives
The GitHub Blog
The GitHub Blog
The Last Watchdog
The Last Watchdog
H
Hackread – Cybersecurity News, Data Breaches, AI and More
博客园 - 【当耐特】
Engineering at Meta
Engineering at Meta
Scott Helme
Scott Helme
Recent Commits to openclaw:main
Recent Commits to openclaw:main
Exploit-DB.com RSS Feed
Exploit-DB.com RSS Feed
量子位
A
Arctic Wolf

Martin Fowler

The Archaeologist’s Copilot DSLs Enable Reliable Use of LLMs Fragments: July 13 Experiences with local models for coding Viability of local models for coding Fragments: July 6 Building Reliable Agentic AI Systems Fragments: June 16 Fragments: June 2 Fragments: May 27 The VibeSec Reckoning bliki: Vibe Coding Maintainability sensors for coding agents Fragments: May 14 bliki: Interrogatory LLM What Is Code? Fragments: May 5 bliki: Mythical Man Month Fragments: April 29 Structured-Prompt-Driven Development (SPDD) Fragments: April 21 Fragments: April 14 Alan Turing play in Cambridge MA Fragments: April 9 Feedback Flywheel Principles of Mechanical Sympathy Harness engineering for coding agent users Fragments: April 2 Encoding Team Standards Fragments: March 26 bliki: Architecture Decision Record Fragments: March 19 Context Anchoring Fragments: March 16 Fragments: March 10 Ideological Resistance to Patents, Followed by Reluctant Pragmatism Ideological Resistance to Patents, Followed by Reluctant Pragmatism Humans and Agents in Software Engineering Loops Design-First Collaboration Fragments: February 25 Knowledge Priming Fragments: February 23 Fragments: February 19 bliki: Host Leadership Fragments: February 18 bliki: Agentic Email bliki: Future Of Software Development My favorite musical discoveries of 2025
bliki: Excessive Bold
Martin Fowler · 2026-01-28 · via Martin Fowler

I'm increasingly seeing a lot of technical and business writing make heavy use of bold font weights, in an attempt to emphasize what the writers think is important. LLMs seem to have picked up and spread this practice widely. But most of this is self-defeating, the more a writer uses typographical emphasis, the less power it has, quickly reaching the point where it loses all its benefits.

There are various typographical tools that are used to emphasize words and phrases, such as: bold, italic, capitals, and underlines. I find that bold is the one that's getting most of the over-use. Using a lot of capitals is rightly reviled as shouting, and when we see it used widely, it raises our doubts on the quality of the underlying thinking. Underlines have become the signal for hyperlinks, so I rarely see this for emphasis any more. Both capitals and underlines have also been seen as rather cheap forms of highlight, since we could do them with typewriters and handwriting, while bold and italics were only possible after the rise of word-processors. (Although I realize most of my readers are too young to remember when word-processors were novel.)

Italics are the subtler form of emphasis. When I use them in a paragraph, they don't leap out to the eye. This allows me to use them in long flows of text when I want to set it apart, and when I use it to emphasize a phrase it only makes its presence felt when I'm fully reading the text. For this reason, I prefer to use italics for emphasis, but I only use it rarely, suggesting it's really important to put stress on the word should I be speaking the paragraph (and I always try to write in the way that I speak).

The greatest value of bold is that draws the eye to the bold text even if the reader isn't reading, but glancing over the page. This is an important property, but one that only works if it's used sparingly. Headings are often done in bold, because the it's important to help the reader navigate a longer document by skimming and looking for headings to find the section I want to read.

I rarely use bold within a prose paragraph, because of my desire to be parsimonious with bold. One use I do like is to highlight unfamiliar words at the point where I explain them. I got this idea from Giarratano and Riley. I noticed that when the unfamiliar term reappeared, I was often unsure what it meant, but glancing back and finding the bold quickly reminded me. The trick here is to place the bold at point of explanation, which is often, but not always, at its first use. 1

1: For example, sometimes a new term will appear first in a list. Eg “We carry out this process in three steps: frobning, gibbling, and eorchisting”. In this case we don't bold the words as they appear in the list but later on when we explain what on earth they mean.

A common idea is to take an important sentence and bold that, so it leaps out while skimming the article. That can be worthwhile, but as ever with this kind of emphasis, its effectiveness is inversely proportional to how often it's used. It's also usually not the best tool for the job. Callouts usually work better. They do a superior job of drawing the eye, and furthermore they don't need to use the same words as in the prose text. This allows me to word the callout better than it could be if it also had to fit in the flow of the prose.

A marginal case is where I see bold used in first clause of each item in a bulleted list. In some ways this is acting like a heading for the text in the list. But we don't need a heading for every paragraph, and the presence of the bullets does enough to draw the eye to the items. And bullet-lists are over used too - I always try to write such things as a prose paragraph instead, as prose flows much better than bullets and is thus more pleasant to read. It's important to write in such a way to make it an enjoyable experience for the reader - even, indeed especially, when I'm also trying to explain things for them.

While writing this, I was tempted to illustrate my point by using excessive bold in a paragraph, showing the problem and hopefully demonstrating why lots of bold loses the power to emphasize and attract the skimming eye. But I also wanted to explain my position clearly, and I felt that illustrating the problem would thus undermine my attempt. So I've confined the example to a final flourish. (And, yes, I have seen text with as much bold as this.)