↓ Ir para o conteúdo principal

← todas as notas

📎 Webclip

Lessons learned in 35 years of making software

Jim Grey says that 35 years in and around software taught him recurring lessons about how work gets done and how careers move forward. He argues for straightforward solutions, small releases to Production, visibility, relationships, and a network beyond the current company.

He also says to be willing to take on unknown tasks, to chase interesting work over salary and title, and to stretch beyond natural tendencies. He adds that white-collar work depends more on relationships and power than on work alone, that perfection often means overdoing it, and that software will eventually be decommissioned anyway.

Reading notes
#

  • He argues for making the simplest possible path, without clever solutions or upfront overengineering.
  • He maintains that software in Production is worth more than long projects waiting for perfection.
  • He says that advancing in a career and seeing a vision become reality depends on relationships within the company.
  • He states that work needs to be visible so that leadership notices what the person can deliver.
  • He reports that a network of contacts outside the company has already helped him recover jobs and get his current role.
  • He says that accepting tasks without knowing how to do them and figuring out the path afterward helps learning and gaining the boss’s confidence.
  • He recommends seeking adventure and interest instead of salary and title, because that has brought variety and growth to his career.
  • He explains that, by tendency, he prefers to work alone and avoid competition, but he had to learn to show value consistently.
  • He notes that, in the white-collar world, relationships and power weigh more than isolated effort, despite the expectation he had from coming from the working class.
  • He states that, when the result seems too ready from a personal point of view, it is likely that it has already gone beyond what the company needed.
  • He says that software ends up being decommissioned over time, so it is worth delivering small increments and iterating instead of overbuilding.