Comments & Suggestions

Stand: 2026-09-10

Comments and suggestions exist in every vault — with or without sync, with or without encryption. In an encrypted workspace they are signed objects (see Security & Sharing); everywhere else they carry the name from Your name (remarks and reviews) in the settings. A suggestion is a comment with a replacement text: both run through the same store, the same column, the same sheet. If the name is missing, Plainva asks for it once at the first comment; without an answer the device signs with its own label (such as “Windows device 4f3a”). Your own remarks read You.

Comments on a note

Commenters get a read-only editor with Comments beside it. Select a passage in the text and write your note next to it — the comment then hangs on that passage rather than on the document. Replies form a thread, Resolve files it away. Comments and resolution markers are encrypted and signed workspace objects themselves; the note stays plain Markdown, and no reply ever lands in its version history. The column shows and hides with the Comments button in the note’s toolbar — it carries the number of open threads; Plainva remembers your choice per vault and otherwise opens the column only when the note has open threads. Its head counts, filters (Open · All) and mutes; every card names who wrote when at the top and shows its actions once you hover or select it. You can delete your own comment or suggestion (trash on the card, the question asked right there); owners and admins can delete anybody’s. Deleted means hidden for everyone — the sealed record stays in the store.

Type an @ in the comment field and Plainva offers the members of the workspace; picking one writes the name into the comment as ordinary text. A thread that names you moves to the top of the column and carries the Mentions you badge — a resolved thread does not, it needs no attention any more. Mentions are never stored; they are read out of the text when it is displayed: rename a member and older comments follow the new name; remove them from the workspace and the highlight falls away while the typed text stays exactly as written.

So a comment finds its passage again after later edits, Plainva wraps it in invisible HTML comments. Other editors do not display them; anyone who would rather not have them in their files turns off Mark commented passages in the note under Content & structure — Plainva then locates the passage through the quoted text. If it can no longer be found, the comment says so instead of pointing somewhere arbitrary. In Live mode these markers are invisible — even on the line you are typing in; the caret steps over them. Only Source mode shows them.

Not every note is about a sentence. A comment can just as well hang on an image, a diagram, a table cell, or a property in the properties panel. Each has its own way in: a small bubble on the image or diagram, an entry in the cell menu, a dot on the property row. On an image you can additionally drag a rectangle to say which part you mean — a click without a drag means the whole picture. The card then names what it hangs on instead of quoting text: “On the image”, “On the marked spot in the image”, “On the diagram”, “Cell · row 2, column “Status””, “On the property “Status””. These anchors write nothing into the note: the marker pair above is for passages, and an image or a cell needs none. A table cell shows the same speech bubble on hover as a picture does; the cell menu (right-click) stays as the second way. A cell that already carries a comment shows a small triangle in its top right corner instead — a click on it opens the card. A cell has one comment; whoever wants to add something replies in its thread. From a database both work directly: a right-click on a cell offers Comment on property (on the phone the entry sits in the sheet a tap on the cell opens), and the small speech bubble with a number in a cell opens the entry right at that remark.

The read view shows the same markings as the editor: tinted passages, commented cells with their corner triangle, and pictures including their marked region; a click opens the card. When several cells of a table are commented, every one of them carries its triangle.

Things move, and the card says so rather than pointing at the wrong place. A cell is found again by what it says: a row inserted above moves it and the card names the new row; a cell that now says something else says exactly that. A property that is gone names the value it held when the comment was written — otherwise the note would still be readable but no longer understandable. A property that was merely renamed is found again through the .base that governs it, so the comment follows the new name. And a commented property shows its dot in the database as well, not only in the note — otherwise a question on a status column would be invisible to everyone looking at the board.

Suggesting changes

Suggest beside Read, Live and Source is the way to propose changes: you type in a copy of the note — deleted text stays struck through, new text is marked — while the file itself does not change. A band above the note counts the changes, takes an optional sentence for the round and offers Discard and Send suggestions. Sending turns every changed block into a proposal; a pure insertion becomes an insertion point that stands between two words. Anyone with comment rights may suggest, write rights or not. The proposals of one send arrive as a round under the Suggestions tab of the comments column: the sentence, the number of blocks and — while more than one is open — Apply all and Decline all. Each block can still be handled on its own with Accept or Decline; accepting rewrites exactly that passage as an ordinary edit, covered by the version history like anything else. If the text has changed in the meantime, nothing is written; you are told the passage no longer matches. The text itself shows an open proposal as Word does: the passage struck through, the new wording right behind it, with Accept and Decline on hover; if you would rather not see that, turn Show changes in the text off in the column head. Leaving the note keeps your unsent copy; opening it again returns you to the mode. Clicking a struck passage or the insertion pill in the text selects the proposal’s card — the column switches to Suggestions and scrolls to it; clicking the card jumps to the passage. Exactly as with comments. If you leave the note, the vault or the app before sending, the copy stays on this device: on the next open, Unsent suggestions found offers to continue or discard — if the note has changed meanwhile, the blocks are found at their places again, and whatever has no place left is appended as a remark.

Make a task turns a thread into work: Plainva creates the task in the Standard task database — with its template, its storage folder and its done column, exactly the way every other task there comes about. The first line of the comment becomes the title, the quoted passage stands as a quote in the task, and a link leads back to the note. The comment itself stays where it is and gains a reply naming the new task — and it still counts as open: the work moved, the conversation did not. With no standard task database set, Plainva says so rather than creating a note somewhere. The reply appears at once, its link to the task is clickable, and the message offers Open; the push to a provider list runs in the background.

On the phone

On the phone the same threads sit in a sheet rather than a column, and tapping the quoted piece jumps to it in the note. Reading, replying, resolving and deciding on suggestions all work there too, and the anchors beyond the running text can be started there as well — an image including a region drawn with your finger, a diagram, a table cell, a property. A property row is opened with a long press there rather than a button: the row is already a button itself, and a button inside it would make it unusable. A comment on a passage starts from the selection in read mode: select the text, then Comment or Suggest in the bar above it. Proposing changes also works from the menu — Suggest in the note’s menu opens the same mode with the same band, and the sheet shows the rounds under Suggestions with Apply all and Decline all. Make a task is there as well. The sheet carries the same head, the same cards, the switch for the inline rendering and deleting; applying and declining stay in the sheet there, not in the text. Tapping a tinted passage or a proposal in the text opens the sheet on the matching card.

Notifications and overview

Plainva can tell you when somebody has written something — Settings → Content & structure → Remark notifications. It stays off until you switch it on, and the moment you do is the zero line: nothing from before is announced afterwards, and it is in the overview anyway. You choose what you are told about: Only mentions, What concerns me (the default — mentions, replies in threads you wrote in, remarks on your notes, and verdicts on your suggestions) or Everything. Two things get through on every setting, because they would otherwise be missed: when a guest of a publication has written back, and when a suggestion is waiting on your decision. Nothing you wrote yourself is ever announced to you. One message per cycle rather than one per remark; only when exactly one is new does it name the note and the person. You silence a single note with the bell in the comment column, and a settled thread never speaks again.

A notification appears on the lock screen, which is why the switch Show note and text in the notification sits right next to it. Turn it off and the message only says that something arrived — no name, no note title, no text. If the vault is locked there is no preview regardless: the remark arrived encrypted, and nothing can open it beforehand. That is not a fault, it is the encryption doing its job. Clicking the message opens the note and highlights the card it means; when several are new it leads to the overview instead. On the phone the message comes later than on the desktop, and that cannot be engineered away: Plainva has no server that could nudge a device — that would mean a foreign server learning when who commented on which note. A remark is therefore noticed where a device looks anyway: in a sync cycle. On the desktop that runs continuously; the phone notices when you open it or when it returns to the foreground.

Open comments gathers every unresolved thread in the vault in one place — otherwise you would have to open each note in turn to see where something is waiting for you. On the desktop you open the overview from the ribbon or the command palette; on the phone it is an area of its own, reachable through Areas and movable into the bar. For me shows only the threads that name you with @ — the same rule as the Mentions you badge —, All shows every open thread. Clicking an entry opens the note at the commented passage. Resolved threads drop out of the list: it is a to-do list, not an archive.

When a gathered notification reports several new remarks, its click opens the overview on the New tab — exactly the threads it meant; All and For me stand next to it.

Where the remarks live

Without a workspace the remarks live in the vault under .plainva/sync/, one file per device, comments.<device>.json — sealed as .enc once a sync passphrase exists. They are never written into the note, and no reply lands in its version history. Every device writes only its own file and reads all of them, which is why nothing collides when the folder is shared through Dropbox, iCloud, Syncthing, a network share or Git: the files travel like any other, and in a Git vault they belong in the history. A Plainva sync connection carries them as well. All devices should run the same Plainva version: an older one reads only the earlier shared file comments.json, which is still read but no longer written.

When another device’s file arrives through a foreign sync, the desktop shows it as soon as the folder reports it; the phone on the next open or on the return to the foreground.

When the vault has a sync passphrase that has not been entered on this device, the remarks are locked: column, sheet and overview say so and offer Unlock. Nothing is read or written until then — a locked device never puts a plaintext file beside the sealed one.

When you rename or move a note (or a whole folder) in Plainva, its remarks follow. Plainva also remembers the move before the first comment arrives or while the older history is locked. After reopening or unlocking, late comments can follow the saved path changes. If you reuse the old filename for a new note, newer remarks on that note stay there.

These local records contain only path changes and their original times. They are kept if a comment file is damaged or cannot be updated; syncing carries the markers in the usual comment bundle, encrypted when a passphrase is active. If the local record itself cannot be saved, the rename still stands and Plainva warns you. Check storage and file access; remarks may remain under the old name in Open comments.

A remarks file that cannot be read is never overwritten: this device’s own file is set aside untouched (comments.<device>.broken-<time>.json), another device’s stays where it is. Plainva reports the reason once and offers Export diagnosis — paths and reasons, never comment text.

Concurrent replies, move markers and an active sync are merged when writing. A missing, empty or older cloud file does not remove remarks already received. When connecting a new sync target, local legacy records and remarks received from other devices are transferred with their original IDs; duplicate copies appear as the same remark.

Before setting aside this device’s damaged file, Plainva reads the complete backup and compares it with the original. If that fails, the original stays and new remarks are not saved; Plainva reports the backup failure. Check free space and file access, then send again. Plaintext sources are removed only when all their content is present in a readable encrypted copy both on this device and at the sync target. A source changed in the meantime stays until the next successful sync.

Incomplete operations and conflicting decisions

Plainva records progress for each comment operation on this device. If the text is saved but the comment update is pending, a notice offers Try again. Retrying continues the unfinished steps without undoing the saved text or later typing. A partially sent suggestion round continues with the same proposals; further changes to your working copy remain unsent.

If the note changed before its text was confirmed, open Compare text. Compare the planned change with the current text and edit the note yourself if needed. Retrying checks again; different text is never replaced with the old version.

If devices report different decisions, the card shows Conflicting decisions. Use Review decision to compare the proposal with the current note. Confirm as applied or Confirm as declined confirms this text and replaces only decisions already known here. It does not insert the proposal again. A previously unknown opposing decision arriving later can show a conflict again.

Older comments in a workspace

Desktop and mobile use the same signed comment store in an encrypted workspace. Comments written offline remain in the outbox. After a sending failure, you can retry the same comment without duplicating a comment already saved.

Older comments in the previous comment files are preserved and imported with workspace management and comment permissions. From earlier history keeps the original name and date. Those claims do not retroactively establish workspace membership; the new signature confirms the import. If permissions, the synced target note or a successful save are missing, the old comment stays visible. Reply, delete and suggestion actions remain unavailable until its target is durably assigned. Importing old history does not send its notifications again.

If only the older history is locked, Unlock opens the sync-passphrase flow; new workspace comments remain available. If the workspace itself is locked, the action opens Security settings. Suggestion operations started before the update keep their saved progress and can resume with current permissions. Importing does not remove the original comment files.