Ideas and entries
A specialist word in your work is an idea. kelve marks each idea. Each idea has an entry from the shared graph that teaches it.
The marks
Each idea’s mark is a small square. In a page read plainly, the square and the idea’s name are in the sentence that uses it. As code, the square is beside the idea’s first line, and its name is in the margin, with a line on how this work uses it.
| Square | What it means |
|---|---|
| Red | New in this change |
| Grey | Already here, or not in this change |
| Filled | You haven’t read its entry yet |
| Hollow | You’ve read it, anywhere |
| Black, on a line | A first principle |
In a file shown as code, press ] and [ to move between marks, and Enter to open one.
An entry
Click a square or a name, and its entry opens beside the work. The work narrows but stays. An entry has:
- The idea’s name and the question it answers.
- Where it came from, “From the shared graph”, with how many sources it was checked against and how often it’s been read.
- The key sentence, then how it works, with examples and diagrams.
- Folds you open in place: a worked example, a common mistake, the precise rule.
- Your notes on the idea, if you left any, with Add a note.
- How it relates: Rests on, Kind of, Answers, Used with, Instead.
- The lessons agents learned using it.
- Its sources, numbered.
Trace it down
Click an idea in an entry, and its entry opens beside the one you were on. Two entries show at once. The path in kelve’s bar has the rest, and each step is a link back. Escape closes the top one.
At the bottom is a first principle. It says “Nothing: this is a first principle, the ground everything above it stands on.”
When there’s no entry yet
If the shared graph has no entry for an idea, your agent writes one when you open it. Until it does, the pane says where the entry is on your agent’s list. Do next moves it up. While the entry is written, the pane shows its steps: Written, Its claims checked, Shown here.
Once the entry is written, it’s in Ready for you, the count in kelve’s bar, until you’ve seen it. You can read it at once. Any claims not checked yet are dotted.
Say you aren’t signed in, and your agent writes an entry the graph didn’t have. The first time, the entry asks whether to add it to the shared graph, or Keep it to myself. Adding to the shared graph says more.
When an entry doesn’t help
Under an entry from the shared graph:
- Didn’t help me understand offers “Needs an example”, “Too hard” and “Missing something”. Only your choice is sent. The next agent that adds to the graph and opens the idea makes the entry better.
- Something wrong? reports a mistake in the entry, or in how it relates to another idea. The shared graph checks it again.
Search
Press ⌘K, or Ctrl K, from anywhere to search. Before you type, search lists:
- Newly written, the entries you asked for in the last day that your agent has written.
- The ideas in what’s open, those in the most files first.
- The lessons about them.
Once you type, ideas come first. Ideas here come before those from the shared graph that aren’t in your work. Opening an idea with no entry keeps it for your agent to write. After the ideas come lessons, files by name, lines that say the words, and notes. Arrows choose, Enter opens, Escape closes.