Your call
The one place that shows what is waiting on you. Decisions need an answer, heads-ups need only a nod, and suggestions from your workspace sit apart so they cannot outrank a person.
What it is
Your call is the one place that shows what is waiting on you: the things someone or something in your workspace needs you to settle, ranked so the most pressing sit at the top.
It exists because the alternative is worse. Work that needs a person tends to scatter across threads, and anything scattered is eventually missed. One place means one thing to check.
Two kinds of item, and they are not the same
Everything in your call is one of two shapes, and the difference matters more than it looks.
A decision needs an answer from you. Someone asked which of two ways to go, or asked you to supply something only you have. A run may be stopped and waiting on it right now, which is why decisions are what the count counts.
A heads-up needs only a nod. Something happened that you should know about. Nothing is being asked and nothing is stuck behind it.
They are drawn differently and counted differently on purpose. A number that counts everything you could look at will always read heavier than the work that actually needs you, and a badge that overstates is a badge you learn to ignore.
If your call says three, three things need an answer. Heads-ups are still there to read; they are just not part of that number.
Suggestions sit apart
Your workspace also proposes things: a rule it thinks would help, a pattern it noticed. These are worth seeing and they are not the same as a person asking you something, so they sit in their own band below rather than in the same ranked order.
The reason is arithmetic. A workspace can produce hundreds of near-identical suggestions from ordinary activity, and if they shared a queue with real questions they would outrank people on volume alone.
Identical suggestions are merged into one row with a count, because a suggestion made two hundred times is one suggestion. Settling any one of them settles the whole group.
The order
- Anything blocking comes first, meaning work is stopped until you answer.
- Then the longest waiting.
- Then your own order, if you have set one.
When there is more waiting than can be counted exactly, you will see a figure with a plus on it. That is a floor, not an estimate: there are at least that many.
Answering
Answer a decision where you find it. You do not open anything first, and you do not go somewhere else to do it, because answering ten things through ten detours is how a queue stops being read.
Answering can also start work rather than only finish it. The first question a new workspace asks you is an ordinary decision in this lane, and answering it opens your first thread, titled with what you typed. If a run was parked waiting on your answer, it picks up where it stopped.
From the command line
The same lane, from a terminal. The command is named register, which is the
internal name for this list:
spl registerChange how many rows print with --limit (30 by default).
Putting one thing above another
When the automatic order is not the order you want to work in:
spl register weight <above> <below>Both arguments are item ids, and a prefix is enough. It persists as your own preference. To take it back:
spl register weight <above> <below> --withdrawTurning your assistant off
Your assistant is the member that answers when you say something to your workspace. It is your preference and you can turn it off:
spl register listener off
spl register listener onSaying one thing
spl say sends one message to your assistant on a fresh thread and prints the
reply. It is the typed equivalent of the first message in the app:
spl say "What changed on the roster today?"| Flag | Effect |
|---|---|
--thread <THREAD> | Continue an existing thread instead of opening a new one |
--wait <SEC> | Seconds to wait for the reply (default 90) |
spl say is one round trip. For work you want to sit with, use
spl run.
Getting back in after time away
spl resumeOne re-entry summary answering three questions: where you were, what changed, and what needs you. Point it at one place to get the summary for just that place:
spl resume th_auth_refactor_abc123
spl resume TASK-0001For scripts
Every command here takes --json (or -o json / -o stream-json):
spl register --json | jq '.data'
spl resume --jsonRelated
- Running a goal: start a run and sit with it.
- Task management: the task lifecycle behind many decisions.
Running work with spl run
Start a run as yourself, watch it stream to your terminal, answer the question it parks on, and read the outcome (with its diff when it changed a repository). How a run is bounded, and how one stops.
Evaluating actors
Score an eval suite against an actor and judge one record over another. The improvement loop scores structure and substance apart, treats a definition's facet as its identity, and compares regressions per axis.