Originally written in French. Translated by AI — the meaning has been preserved, not the prose.
Main idea
PARA is oriented toward action: a project creates a need, and the available knowledge is mobilized to produce a result. Zettelkasten leaves room for emergence: readings become notes, notes enter into relation, and new ideas appear in the network.
A core of pillar cards occupies neither of those places. It is neither a general base nor a project management method: it's a limited set of ideas a person wants to keep active inside themselves, and takes out to think through a tension of the moment.
The three therefore coexist without competing. Asking which one replaces the others amounts to comparing setups that don't answer the same question: what to do now, what emerges from the network, what I want to have in mind.
Layer added by "I went looking for what I had really invented in "Contexte"" (2026-09-21). A fourth setup joins the comparison, discovered long after pillar cards were put in place: Jirō Kawakita's KJ法, where field data written on freely movable cards is brought together and recombined until a structure emerges. As early as 1969, Tadao Umesao described a neighbouring system, where the knowledge born of recombination re-enters the stock. The kinship of gesture is real — getting units of thought out of your head, making them manipulable, bringing them together to keep thinking — but the purposes diverge: the KJ法 starts from scattered material to bring out a structure, pillar cards keep a few already stabilized ideas active. A fourth distinct need, then, and not an ancestor.
Why it matters
It defuses a badly framed question — which knowledge management method to adopt — by replacing it with an inventory of genuinely distinct needs.
It also avoids a design confusion: applying the rules of a note base to a core of convictions turns it back into a note base.
Nuances and limits
Three setups mean three upkeep costs. Coexistence only works if each stays in its role, failing which redundancy outweighs the gain.
Open questions
How, in practice, do entries divide among the three — given that a single reading can feed all three setups at once?