Poul-Henning Kamp recounts how a FreeBSD mailing-list dispute in 1999 gave the computing world the term “bikeshed”. At the time, he had become a de facto spiritual leader of the FreeBSD project and was writing messages about issues he considered serious. One especially heated discussion took him several days to process before he could describe the broader pattern clearly.

The immediate trigger was a proposal for sleep(1) to accept fractional seconds. The change would have preserved existing behavior and aligned FreeBSD with OpenBSD and NetBSD. A long thread produced a disproportionate number of objections and suggestions, even though the patch was small. Kamp respected the original contributor for continuing until the change entered the FreeBSD tree.

Kamp connects the episode to C. Northcote Parkinson’s book “Parkinson’s Law”, published in the early 1960s. Parkinson’s example contrasts a complex atomic power plant with a simple bike shed. People struggle to assess the large project and assume that experts handled its details. A small shed feels easy enough for everyone to judge, so it attracts personal preferences and visible claims of involvement. Kamp calls this impulse “setting your fingerprint”, a phrase used in Denmark for the same form of personal pride and prestige. He also cites Richard P. Feynman’s accounts of Los Alamos as a related example.

The email later became a FreeBSD term of art, effectively a yellow card for an argument that consumes too much attention. FreeBSD committers carried the expression to other projects. Ben and Brian from Subversion then used the idea in their talk “How Open Source Projects Survive Poisonous People” and created bikeshed.org with Kamp’s email. Kamp sees that work as an important point in FOSS history because it offered practical management strategies for open-source projects. Earlier discussions had largely settled for the “herding cats” metaphor.

Kamp’s former core.0 colleagues told him that the talk described problems they had already experienced during the previous decade. They had never presented that management knowledge to other projects. He credits Ben and Brian with making the human side of FOSS project management discussable and respectable. A photograph from BSDCan2007 shows Kamp and Warner wearing the original BSDCon’03 anti-bikeshed shirts beside Ben and Brian.

For BSDCon’03, Kamp designed those shirts as a gesture of support for the project team. A local embroidery shop produced them. He only later realized that each shirt could have shown sheds in different colors at no additional cost. The shirts became popular and remained available through CafePress and other online stores.

Kamp adds a second management lesson, which he calls the “Goodbye & Good Luck” force multiplier. Open-source projects usually try to preserve unity when strategic disagreements appear. He argues that a split can sometimes give both sides more room to build. His example is OpenBSD, which emerged after Theo de Raadt left NetBSD and later produced major security software such as OpenSSH. Kamp had hoped for a similar outcome after FreeBSD removed Matt Dillon, while judging DragonFlyBSD’s progress disappointing at that point.

A later Google search feature produced the phrase “People also search for” around Kamp’s name and led to an amusing dinner-party suggestion: “I blame… The bikshed email”. Kamp then republishes the original message. It was sent by Poul-Henning Kamp from phk@freebsd.org on Sat, 02 Oct 1999 16:14:10 +0200. Its Message-ID was <18238.938873650@critter.freebsd.dk>, and it was blind-copied to the committers and hackers lists.

In that email, Kamp says the sleep(1) thread had already lasted many days. He uses it to explain why a small, well-prepared change can trigger intense participation. His management source is Parkinson’s account of people approving billion-dollar infrastructure while debating a bike shed’s details. The small object offers an easy place to display judgment, ownership, and status.

Kamp also addresses mailing-list noise. He had left the hackers list years earlier because of its volume and had abandoned other lists for the same reason. Filters sent messages from people such as Brett Glass to /dev/null, along with commits concerning unfamiliar document languages and ports. The remaining traffic still overwhelmed him.

He proposes two broad goals: reduce unnecessary list traffic and give contributors room to complete small tasks without hostile intervention. His imagined mail-program warnings would tell senders that several hundred thousand people might spend at least ten seconds reading a message, consuming at least two person-weeks in total, with some recipients paying for the download. Other warnings would note that the sender had not read the whole thread or message. A one-hour cool-off period would block an immediate reply. A further warning would compare the typing rate, shown as N.NN cps, with a slower reference rate, shown as A.AA cps, and flag the likely emotional or incoherent result.

Kamp attributes some of the hostility to reactionary conservatism among older contributors. He offers no technical remedy. His advice is to avoid feeding disruptive voices and to help new contributors survive the mailing-list culture. He apologizes to people who were discouraged from participating and says that the project should provide a broader and friendlier base for contributors.