These are the next seeds the community could build, in order. Each one stands on the ones before it. We'll work on three or four at a time.
There are no dates. A seed is done when it's useful and it has been through an open comment round, not when a calendar says so.
Want to lead one? Say so in the "Governance" discussion category.
Open work on the seed paper
The seed paper is a working draft. This is what we know still needs doing, so nobody has to guess.
- Fold in the newest sources. Version 7.9 lists nine new sources as A1 to A9, after the references. The next full version will number them in order of first appearance, with the rest.
- Ideas still under discussion. A review on 10 October, done with the help of an AI model, raised ideas the paper doesn't settle yet. Three are open threads: AI demand declared as a policy, whether a market price changes the coefficient, and a shared dataset of declarations.
- The method is untested. Section 13 says what hasn't been measured. The tests in Section 9.2 have not been run anywhere we know of.
Say in the "Governance" category if you'd like to take one on.
1. A shared glossary
What. One page of terms, each with a one-sentence definition, synonyms and "see also" pointers. Versioned on its own, separately from the rest.
Why first. Every other seed uses these words. Planners in different companies use the same word for different things ("commit", "allocation", "buffer"), and in our experience that confusion costs real money in real meetings. A glossary can be cited without anyone having to agree with the rest of the body of knowledge.
Starts from. The terms in the seed paper, plus the terms in public sources it cites.
A first version looks like. A few dozen core terms, each through the open period, with the common confusions written down.
Help today. Open a "Term" discussion, or answer the first LinkedIn prompt.
2. A case library
What. Short, anonymized "what happened" stories. Each one gives the setting, the plan, what happened, why, and the lesson, and links to the sections it illustrates.
Why. Planners learn most from misses, and as far as we can find, few are public. A library of cases with the names and numbers changed teaches the shape of a problem without exposing anyone.
A first version looks like. A dozen cases from different settings: a large fleet, an AI cluster, cloud commitments, a mid-sized company, a supply delay.
Help today. Use the "Share a case" form. Read the anonymizing checklist in CONTRIBUTING.md first.
3. A capacity planner competency map
What. What a capacity planner needs to know and do, at a few levels: learning, practicing, leading. Areas might include demand and forecasting, supply and lead times, cost and coefficients, contracts, finance, operating cadence, and the people work of brokering an agreement.
Why. In our experience, many new planners come in from hardware, accelerators or infrastructure engineering. They know the machines well. The map shows what else the job asks, and where to learn it. It also helps managers hire and grow planners.
A first version looks like. One page: areas down the side, levels across the top, a sentence in each cell, and links into the body of knowledge and the reading list.
Help today. Answer the second LinkedIn prompt, or open a "Proposals" discussion with the areas you'd include.
4. Worksheets as tools
What. Printable, fill-in worksheets for the methods, with a worked example in illustrative numbers. The seed paper's Appendix A, a worksheet for one declaration, is the first one.
Why. A method nobody can run on Monday doesn't help much. Worksheets turn ideas into something a planner can try, test and report back on.
A first version looks like. Three or four worksheets, each tried by at least one person outside the editors, with their feedback recorded.
Help today. Try the Appendix A worksheet on an invented case and say where it got stuck.
5. A reading list
What. Public sources worth reading, grouped by topic and by reader: new planner, experienced planner, finance, supply. Each with one line on why it's worth the time.
Why. The good material is scattered across books, papers, talks and blog posts. It includes CMG's work, ITIL, the SRE books, Hixson and Guliani's 2015 article, and the books by Gunther and Allspaw. A curated list saves every new planner the same search.
A first version looks like. The seed paper's references, sorted and annotated, plus what contributors add.
Help today. Use the "Add a public source" form. This is the easiest seed to contribute to.
6. A survey of practitioners
What. A short, anonymous survey of people who do capacity planning. It asks what they own, what they call things, and which tools and methods they use. It also asks where plans go wrong, and what they'd want from a body of knowledge.
Why. Edition 1 should rest on what practitioners actually do, not on what one paper argues. A survey is the simplest way to check one author's view against many people's practice.
A first version looks like. A survey drafted in the open, piloted with a few planners, then shared through the group and partner communities. Results published in aggregate under CC BY-SA 4.0. Raw answers never published if they could identify anyone.
Help today. Suggest questions in a "Proposals" discussion. Say which questions you'd answer, and which you wouldn't.
After these
Edition 1. It comes after the survey and an open comment round on a full draft, and ships with a "what changed" table and a contributor list. See CHARTER.md.