简体中文
EnglishWebsite review
This review may include AI-generated content and is not individually verified. Domain, certificate and other facts come from the observed data on this page.
What is The Third Bit?
The Third Bit is Greg Wilson's personal site about software design, research software engineering and teaching. Rather than a blog about news or opinion, it functions as a hub for his books, workshops, papers and ongoing writing, with a strong emphasis on practical craft: how to build tools, how to manage research software teams and how to teach technical material well.
What you'll find there
- Books — titles such as Software Design by Example in Python and its JavaScript counterpart, Research Software Engineering with Python, Teaching Tech Together and The Architecture of Open Source Applications. The "design by example" books are the most distinctive: they teach design by walking through working implementations of everyday programmer tools.
- Workshops — one-day sessions on organizational change, winding down projects, teaching practices for programmers and data scientists, evaluating AI studies, and managing research software teams, offered in person or online.
- Recent posts — a running series of short, technical entries, for example a sequence on a small language called Frml covering loops, arrays, built-in functions and contracts.
- Papers — short, practical pieces such as "Ten simple rules for writing a technical book" and "Twelve quick tips for software design."
Who it suits
The site is most useful to three kinds of reader: research software engineers who need to build and maintain code in an academic setting; programmers and data scientists who teach or run training; and developers who want to learn design by reading complete, working code rather than abstract principles. Someone looking for quick how-tos on a specific framework will find less here than someone interested in the reasoning behind how tools are structured.
A practical next step
If you're deciding where to start, match the resource to your situation. New to research software? Begin with Research Software Engineering with Python. Comfortable coding but shaky on design? Pick the Software Design by Example edition in the language you actually use. Facing a teaching or team-leadership problem? The workshop list is the fastest way to see whether a topic matches, and the contact route for hosting is stated on the page. For related reading, Software Carpentry covers the training organization Wilson co-founded, and pyOpenSci offers a community angle on research Python practice.
Who is Greg Wilson and what is his background in software education?
Greg Wilson is a software educator, author and researcher best known for founding Software Carpentry, the volunteer organization that teaches computing skills to researchers. His site The Third Bit describes him as co-founder and first Executive Director of Software Carpentry, co-founder of It Will Never Work in Theory (which translates software engineering research for practitioners) and co-founder of The Architecture of Open Source Applications. He holds a Ph.D. in Computer Science from the University of Edinburgh, has been named ACM SIGSOFT Influential Educator of the Year, is a Fellow of the Python Software Foundation, and has co-won a Jolt Award. He has authored, co-authored or edited 14 books on programming, plus two for children.
What the teaching background looks like in practice
His education work runs along two tracks. One is teaching researchers and scientists to program well enough to trust their own results — the Software Carpentry model of short, hands-on workshops. The other is teaching programmers how to teach: Teaching Tech Together (2019) is an evidence-based introduction to instruction for people with technical backgrounds, and the site lists a one-day workshop of the same name.
The books show the same preference for learning by building. Software Design by Example in Python (2024) and its JavaScript edition (2022) each work through a couple of dozen small programs that reimplement tools programmers use daily, so design decisions become concrete rather than abstract. Research Software Engineering with Python (2021) extends this to running research software projects and teams.
Workshops as a window into his range
The workshop list on the site is a useful summary of what he considers adjacent to software education: organizational change for people with research backgrounds, winding down projects deliberately or on short notice, managing research software teams, and "How to Not Be Wrong About AI" — telling good studies from bad ones. If you are choosing a speaker or trainer, that list is the fastest way to judge fit.
A practical next step
If you want to sample his teaching before committing to a book, start with the Software Design by Example edition in the language you actually use, and read the most recent posts (the site shows a series on a small language called Frml) to see how he explains a design incrementally. For a team or department, the workshop descriptions indicate whether the need is craft, management or teaching skill. Related communities worth knowing include The Carpentries, the organization that grew out of Software Carpentry.
Which of Greg Wilson's books should I start with for learning software design?
Start with Software Design by Example in Python if your goal is learning software design through worked examples. It is the most direct fit: the site describes it as two dozen examples that teach design by building the tools programmers use every day, and the Python edition is the newer of the two language versions. Choose the JavaScript edition instead if JavaScript is the language you work in daily; the approach is the same, with twenty examples.
How the books differ
| Book | Best for | Trade-off |
|---|---|---|
| Software Design by Example in Python | Learners who want many small, buildable design examples in Python | Python-specific; you must be comfortable reading and running Python code |
| Software Design by Example in JavaScript | JavaScript developers who want the same example-driven approach | Fewer examples than the Python edition |
| Research Software Engineering with Python | People building and running research software projects, including team and process concerns | Broader than design; less focused on the craft of structuring individual programs |
| Teaching Tech Together | Programmers and data scientists who teach or train others | About teaching practice, not software design |
A practical next step: pick one example from the Python book that resembles a tool you already use, such as a parser or a small command-line utility, and rebuild it before reading the chapter's explanation. If the code feels foreign, switch to the JavaScript edition to stay in a familiar language; if you find yourself more interested in how to run a team around such code, move on to Research Software Engineering with Python. The site also lists workshops, including "Teaching Tech Together" and "Managing Research Software Projects," which suit readers who want the material delivered in person rather than through a book.
For related free reading on design and computing craft, The Third Bit is the author's own site, where the book descriptions and recent posts appear.
How can I host one of the listed workshops in person or online?
Email Greg Wilson directly using the contact address on The Third Bit and name the workshop you want to host; the site states that hosting requests, whether in person or online, go through email, and points to a rider for details.
What to include in your request
- The workshop title, copied from the site's list (for example, "Teaching Tech Together," "Organizational Change," "Closing Time," "Managing Research Software Projects," or "How to Not Be Wrong About AI").
- Whether you want it in person or online, plus your location and time zone.
- Your audience: roughly how many people, their backgrounds (researchers, programmers, data scientists, research software developers), and what you want them to leave with.
- Possible dates or a date window, and the format you can support (one day is the length given for "Teaching Tech Together"; confirm the others).
- Who will handle logistics such as the venue, invitations, and recording, if any.
Choosing the right workshop
| Audience need | Workshop to ask about |
|---|---|
| Researchers facing institutional or process change | Organizational Change |
| A team ending a research or software project | Closing Time |
| Programmers or data scientists new to teaching | Teaching Tech Together |
| Engineers evaluating AI claims and studies | How to Not Be Wrong About AI |
| Leads of research software teams | Managing Research Software Projects |
As a next step, read the rider linked from the workshops section before you write, so your email already addresses the conditions for hosting. If your group is closer to self-study, the site's books, such as Teaching Tech Together and Research Software Engineering with Python, cover similar ground and can help you decide whether a hosted session is worth requesting.
What does the 'How to Not Be Wrong About AI' workshop teach software engineers?
It teaches software engineers how to read AI research critically—specifically, how to tell credible studies from weak ones. The workshop is framed around what engineers need to understand in order to judge the quality of evidence presented about AI, rather than around building or deploying AI systems. The stated purpose is helping participants "tell good studies from bad ones," which points to skills in evaluating methodology, claims and the gap between reported results and real-world conclusions.
Who it's for and what to expect
The audience is software engineers, so the material assumes technical comfort but not necessarily statistics or research-methods training. Expect the emphasis to fall on practical judgment: what questions to ask about a study, which claims are supported by the evidence shown, and where enthusiasm for a result outruns what was actually demonstrated. That makes it most useful to engineers who are being asked to adopt AI tools on the strength of vendor claims or headline findings, or who need to push back on decisions justified by shaky evidence.
A concrete use
Suppose your team is weighing an AI coding assistant because a paper reports large productivity gains. The workshop's framing suggests the useful questions: what was measured, on whom, over how long, against what baseline, and whether the tasks resemble your work. That is a more defensible way to decide than adopting on the strength of the abstract.
Practical next step
If you want to host it, the site invites email inquiries for in-person or online sessions and points to a rider with details. For self-study in the meantime, Greg Wilson's writing on evidence and teaching—such as Teaching Tech Together—covers related ground on evaluating claims, and The Third Bit lists his posts and publications for further reading.
Where can I find Greg Wilson's recent posts and publications?
Greg Wilson's own site, The Third Bit, is the most direct place to find both. The homepage lists his "Recent Posts" as a dated table (for example, a run of entries on "Frml" in October 2026) with a link to read them all, and a "Recent Papers" table covering 2020–2024, alongside a "Books" section describing his titles.
What each section gives you
- Recent Posts: short, dated entries, best for following current work and thinking as it develops.
- Recent Papers: a year-by-year list of papers, useful if you need citable work such as "Ten simple rules for writing a technical book" (2023) or "Twelve quick tips for software design" (2022).
- Books: descriptions of his longer works, including Software Design by Example in Python (2024), Software Design by Example in JavaScript (2022), and Teaching Tech Together (2019).
- Workshops: a separate list of one-day sessions (organizational change, teaching practices, managing research software projects) if you want to host or attend one.
The site also points to a Mastodon account for following posts, which suits readers who prefer updates in a feed rather than checking the page.
A practical next step
If you are a researcher or developer deciding what to read first: use the Posts table to gauge current interests, then pick a book matching your language—Python or JavaScript—since the two Software Design by Example volumes teach the same approach through different code. For teaching or training roles, Teaching Tech Together is the natural starting point.
Fastly
User reviews (0)