Updates

  1. Person pages list letters, writing, profiles and links as sortable tables with filters, and show each person's companies as links; Fund letters show each letter's description in its own column.
  2. Experts show a checked job title and a short description of each person; 30 more people have photographs.
  3. Every company, fund, university and publication the experts work for has a page listing the people there.
  4. Fund letters show each letter's thesis and filter by it; 225 person pages list the essays, columns and newsletter issues the person wrote.
  5. Person pages list every essay and newsletter issue a person wrote, and their own books as a table in List view.
  6. Recommends shows the five largest kinds of book and the rest under More; browse pictures load sooner.
  7. Podcasts and talks link the original episode, filter by Audio and video or Press, and list each publisher once.
  8. Pages are wider, Recommends has List and Grid, and a filter clicked again turns off.
  9. Book cards sit in a ruled grid with Open Library and Goodreads links and each book's one-line note.
  10. Funds is now Investors; each investor's styles are buttons that filter the table.
  11. Podcasts and talks on a person's page filter by year and publisher and show as a list or a grid of pictures.
  12. Quotes show only people's own words: 472 summaries written by list sites are gone.
  13. Country pages open from Browse again, and every link on the shelf lands on a page.
  14. Subject and person pages show their sections as tabs, and each book appears once on a person's page.
  15. Book cards show the quote, who said it and the source; grids show five cards to a row.
  16. On a phone, a menu holds the sections, search, Updates, Feedback and dark mode.
  17. Feedback takes a screenshot of the page with no browser prompt, opens a screenshot full size when you click it, and clears itself once sent.
  18. Book cards are centred, search names the topic each book sits in, and any text cut short shows whole on hover.
  19. Search groups its results by kind: people, companies, geographies, subjects, topics, books, funds, fund letters and papers.
  20. The Books table lists who recommended each book by name, and the browser tab names the section you are in.
  21. Feedback can take a screenshot of the page, or one you paste or choose, and opens in the corner beside Updates.
  22. Each rail button reopens its section where you left it, and Browse, Table and People show on every book and subject page.
  23. Filters grey out choices that hold nothing instead of hiding them, and Books › People filters by authors or recommenders.
  24. Feedback takes a subject line and up to three screenshots.
  25. Biographies are grouped by the person they are about: 41 new pages, from Abraham Lincoln to Lee Kuan Yew.
  26. Sport splits into 23 sports, and Religion into one topic per faith.
  27. Book lists show as cards: the cover, one recommendation quote and every recommender by name.
  28. Seventeen subjects that had two or three pages under different names now have one page each.
  29. Every page lists its Experts as authors and recommenders, with a link to them on the Experts tab.
  30. Economics, Financial history, Financial panics and Accounting are subjects of their own, with one page per financial crisis.
  31. Book, person and subject pages open inside the app, so the tabs stay in view.
  32. 996 more book covers, and photographs or logos for 227 more people and companies.
  33. Books › Browse shows every subject as cards, and the rail holds Fund letters, Papers, Experts, Books and Funds.
  34. Every topic over 150 books splits into narrower topics, each with a page of its own.

Send feedback

What is it?
Screenshots, up to 3 (optional)
or paste one with Ctrl+V

We use your email only to reply to you.

The Soul of a New Machine

The Soul of a New Machine

Tracy Kidder
1981
T1 293 pages · about 4h 53m to read Open Librarygoodreads

Pulitzer-winning account of building a minicomputer against the clock

Summary

Kidder embedded with a Data General engineering team in 1978 and 1979 as they designed a thirty-two bit minicomputer, the Eagle, in about a year with a small team, few resources and no official sanction at the start. The book won the Pulitzer Prize and remains the standard account of what engineering work actually involves.

The organisational situation drives the narrative. Data General had committed its resources to a different project in North Carolina, and Tom West's team in Massachusetts built a competing machine in parallel, initially concealed. West recruited young engineers straight from university, gave them problems beyond their experience, and used what the team called signing up, meaning a voluntary commitment to unlimited hours in exchange for the responsibility.

Kidder describes the technical work accurately enough to be followed by a non-specialist, covering microcode, the debugging process, the logic analyser and the physical assembly. He also describes the psychological structure: the pinball metaphor in which winning means being allowed to play again, the deliberate withholding of information by West to preserve the team's sense of ownership, and the exhaustion and departures that followed the project's completion.

The book is a period document. Minicomputers no longer exist, and the specific technology is obsolete. The organisational pattern it describes, in which a small unofficial team outbuilds a resourced official one, recurs constantly. The treatment of motivation, in which the reward for shipping is the chance to do it again, describes something that most management literature does not capture. There are almost no women in the account, which reflects the industry at that time.

Key takeaways 8

  1. A small team working without official sanction built a competing product faster than the resourced official project inside the same company.
  2. Recruiting inexperienced engineers and giving them responsibility beyond their experience produced work that experienced engineers would have declined to attempt.
  3. The motivation was the work itself and the chance to do it again, rather than compensation, and the team's term for the reward was being allowed to play the next game.
  4. West deliberately withheld information from his team to preserve their sense that the problems were theirs to solve.
  5. Signing up, meaning a voluntary commitment to unlimited hours in exchange for real responsibility, was the recruitment mechanism and was understood as such by both sides.
  6. The project consumed the people who built it, and most of the team left the company shortly afterwards.
  7. Debugging consumed more of the schedule than design, and the ability to locate faults determined whether the machine would ship.
  8. Internal competition between two teams in the same company produced the winning product, at the cost of duplicated effort and organisational conflict.

Who should read it

Anyone managing engineering teams who wants an honest account of what motivates technical work and what an intense project costs the people in it.

Who recommended it 6 · 2 direct

“Must-read for anyone in a product development or a startup environment.”

Brad Feldreported
Fred Wilsonreported
Paul Grahamreported

Similar books 6

Chosen by shared topics, the entries that hold them, and the people who recommended both.

Walter Isaacson · 2014

Computing as a run of collaborations, from Ada Lovelace and Babbage to Google.

Steven Levy · 1984

Three generations of hackers, from MIT's AI Lab to the Homebrew club and game makers.

Lynn White · 1962

How the stirrup and the plough reorganised medieval society and economy

Recommended by: Patrick Collison, Paul Graham

Harold Abelson, Gerald Jay Sussman · 1985

MIT's introductory computer science text, taught through Lisp and Scheme.

Clayton Christensen · 1997

Why well-run firms lose to cheaper, worse products that enter at the market's bottom.

“Awesome book. May move soon to my Top Ten List. If anything will keep CEOs awake at night, reading this book certainly will.” — Mohnish Pabrai

Source: Mohnish's Bookshelf — Books on Sales & Marketing, entry 66

Recommended by: Jim Collins, Mohnish Pabrai, Aaron Levie and 27 others

Frederick P. Brooks · 1974

Lessons from IBM's OS/360: why adding people to a late software project slows it.

Where it sits in the reference 3

Business themes
Business themes
Industries

Topics

Technology and ComputingHistory of computing and semiconductors