QQivo

The task window

Opening a task floats one window over a blurred backdrop. Its first line is a breadcrumb — project → sub-project, saying where the task lives — with the window's own controls in the top-right corner beside it: the subscribe eye (§11 — open when you are subscribed, struck through when you are not; click it to change your mind), the permanent ID, the "…" actions menu — Copy task ID, Copy link (permanent), Move to sub-project…, Archive, Delete (two-click) — and Close. The task title (click it to edit in place) sits on the line below, under the chrome that acts on it. Under that sit two columns: the discussion (left, primary) and the details (right); when there isn't room for both they stack inside the same window. It is also the Inbox's right pane (§11), where it fills the half beside your messages instead of floating — no backdrop, and no room for the details column beside them, so there the window is the discussion alone.

✦ What sets this apart — The layout is discussion-first: the conversation is the wide column and the description is pinned inside it, while the metadata stands aside. Mainstream tools bury the discussion at the bottom of a form.

The discussion column#

  • The description is pinned on top and never scrolls away. It's a true WYSIWYG editor — no view/edit toggle, no raw markdown ever shown — with headings, bold/italic/strikethrough/underline, inline code and code blocks, quotes, lists, links (Ctrl/Cmd+K), and markdown-as-you-type. Type @ to mention a teammate: a user picker opens under the caret (arrow keys + Enter, or click), the mention renders as a highlighted chip, and — in descriptions and comments alike — the mentioned person gets a message in their Inbox (§11), unless they wrote the text themselves, and starts following the task from then on. A formatting row sits above the text in every editor — text style, B/I/S/U, link, quote, code, lists (Ctrl/Cmd+B/I/U/E, Ctrl/Cmd+Shift+X) — applying to the selection, or to what you type next when there is none; it stays put while the field scrolls. Clicking a link pops Open / Edit / Remove. Paste or drag images straight into the text — they render in place with hover Copy/Download, double-click for full size (and an image can never be pasted across tasks). Saving is explicit: a Save / Discard bar appears the moment your draft differs (Ctrl/Cmd+S) — and it stays pinned to the bottom of the field, so a long draft never scrolls its own buttons out of sight. The field scrolls inside itself rather than growing without end, the line you are typing always stays clear of both the formatting row and the buttons, and a field that would otherwise slide below the edge of the panel brings itself back into view.
  • Comments use the same rich editor (text-only — images belong in attachments), and the box sits at the end of the conversation, directly under the newest message rather than parked at the foot of the panel. Above it, comments and system activity interleave on one spine that reads like a chat: oldest at the top, newest at the bottom. It opens scrolled to the end, so the latest word is the first thing you see. When a task has news you haven't read, a New line marks where it starts — see §11. There is an All / Comments filter. Comments — just the conversation — is the default; flip to All for the task's whole life in one feed: created, moved, assigned, blocked, rescheduled, labeled, linked, attached. Field edits narrate both sides of the change — "moved this task from To Do to In Progress", "changed the remaining time on this task from 80 h to 40 h", "updated the description of this task from “…” to “…”" — so the feed answers what it was before without archaeology. (Description excerpts are capped at ~80 characters a side, and the blocked flag logs as plain "blocked"/"unblocked".)
  • You can edit or delete your own comments; a project lead can moderate anyone's — and an edited comment discloses it in place: a pen glyph shows when it was last edited and by whom, marked "(project lead)" when that wasn't the author.
  • Either half can be read on its own. The Description and Activity headings each carry a pop-out button, and pressing one opens that section — and nothing else — in its own window: a long specification with the full height to read (and write) it in, or a long thread without the description above it. It opens under the same heading the task window wears — the project → sub-project breadcrumb, the task's name below it, and its permanent ID in the corner — so there is never any doubt which task you are reading half of; the section's own heading follows underneath. That heading is a statement here rather than a set of controls: the breadcrumb doesn't lead anywhere, the name can't be edited, and the subscribe eye isn't there either, because this is a view of the task, not the window it lives in — what acts on the task stays one Escape behind. The window is deliberately a size smaller than the task window on every side, so it is plain that something new opened rather than the task having changed shape, and it closes on the X in its corner or on Esc, which touches nothing else. The section moves there rather than being copied, so there is only ever one draft of a description and one reply box, and the All / Comments filter is the same switch in both places.

✦ What sets this apart — Every edit, description included, narrates both sides of the change right in the feed; and lead moderation that is disclosed right on the comment instead of happening silently.

The details column#

Status, priority, creator, assignee, Remaining hours (an input on a leaf task; on a task with subtasks, a read-only Σ sum over them — see §2), Due (with the delay chip, §7), and the Plan pickers — always visible; picking a date on an unscheduled task schedules it. Where your access is read-only (§12) each of them still shows its value but will not open: the control is disabled and says why rather than sitting there live and having every change undone a moment later. The planned end can never pass the due date: an offending pick is rejected with a toast, not silently clamped. Then labels (press L), attachments (drag-and-drop, with the team owning team's size cap), progress over subtasks, and the Parent / Subtasks / Linked tasks sections — each with a search picker that can also create the related task on the spot when nothing matches. (The "…" actions menu lives in the window header, top right.)

Moving a task#

The Move dialog lists every sub-project you can write in inside the task's own organization — across projects and teams. Comments, attachments, labels, activity and relations all survive; the ID never changes. Labels are an organization-wide vocabulary, so a task keeps them wherever in the organization it lands — the dialog says so, and there is no warning left to give.

Archiving — Done work leaves the room on its own#

Archived is a second axis, not a sixth status: a task keeps its workflow status (Backlog … Done) and additionally becomes archived — off the board, roadmap, overview, filters and workload math, but never deleted.

  • Automatic: a task that has sat in Done for the owning team's auto-archive threshold (default 30 days, configurable 1–3650 in the team's settings) is archived by a nightly sweep. The sweep works leaf-up — a Done parent waits until its subtasks have been archived themselves, so it never drags unfinished work along. Finished work fades out of the working set by itself — nobody has to tidy the board.
  • By hand: the task window's "…" menu archives any task you can edit. Archiving takes the whole subtask tree with it, so the rule follows the subtree: a task whose subtree is entirely Done archives on the spot; anything that would take unfinished work off the board — the task itself, or any of its subtasks — first asks why, and the dialog lists every subtask that goes along (key, title and status), so nothing is archived sight unseen. The note you type is required and lands in the task's comments, so the reason survives for whoever finds the task later. (The two machine surfaces, §13, archive without collecting a reason — automation states its reasons in its own systems.)
  • The Archive page (Archived tasks in a project or sub-project row's "…" menu, which opens it pre-scoped to that project) is a separate surface, loaded only when you open it and only for one project at a time: pick a project (or one sub-project), search within its archived tasks with the same multi-word filter as the board, and Restore any row — the task returns to its board immediately, parents included if they were archived too. Restoring a still-Done task restarts its auto-archive clock, so a restore always buys the full threshold again. Archived tasks are never loaded anywhere else — a shared task link only resolves again once the task is restored.

✦ What sets this apart — Boards stay honest without housekeeping: finished work retires itself on your schedule, archiving unfinished work in the app demands a written reason that sticks to the task, and nothing archived is more than two clicks from coming back.