Screens that change themselves

Most of what a screen needs to do by itself is not a cue. A countdown should retire when the speaker appears. A card should turn red in the last five minutes. A message should fade in at 18:00 and a clock should dim at night. None of that is worth an operator, and none of it should need one.

A rule says: while these conditions hold, this block looks like that. It never edits the block — when the conditions stop holding, the block is itself again, so there is nothing to undo and nothing to remember to put back.

Open a scene in the shot editor, select a block, and use the Rules tab beside Content and Design. A rule on the canvas changes the background instead — same grammar, set on the shot or the layout.

Start from a preset — Appear at a time, Dimmer at night, Only until the speaker is on, Red when the talk overruns — then adjust it. Each rule tells you whether it holds now, on this device's clock.

Each sensor reads one thing. Pick the sensor, then which of its readings you mean, then how to compare it.

• Time of day — from, before, between, or not between two times.
• Time of day you named — the workspace's dayparts, so "evening" means one thing everywhere.
• Weekday and date — on, from, until, between.
• A moment in the data — "within 10 minutes before {starts_at}".
• A data value — is, is not, contains, empty.
• The playhead — where the shot has got to, for a block that should arrive two seconds in.
• The live session — where it is, when it starts, how long it has been on air, whether it has overrun.
• Someone in front of the screen — read by the screen's camera: someone is there, nobody is there, someone has been there a while. On a block, or on the stage to switch the whole story (People in front of the screen).
• A named condition — one you defined yourself (below).

The session sensor is what lets one slide carry a whole arc. Put the topic, a countdown and the presenter block in one shot, then add two rules:

• the topic card and countdown hide once the session is live;
• the lower third fades 20 seconds after it went on air.

That is the sequence a room actually wants, and it needs no cue, no show and nobody watching. On-air time counts from when the talk actually started, so a speaker who begins ten minutes late still gets their full lower third.

A rule holds when all of its conditions hold, or when any one does. For anything more exact, add a group: a bracket whose conditions are judged together and then count as one.

After 18:00, on a Friday or a Saturday is a clock condition and a group containing the two weekdays. Without the bracket you would be asking for Friday and Saturday — never true — or for anything after 18:00 or any Friday, which is most of the week. The sentence under the panel always reads back what you actually built.

When the same meaning turns up in more than one rule, give it a name once under Settings → Named conditions: between_talks, doors_open, last_five_minutes. Any rule can then ask for it by name.

The point is that it has one definition. Change what "between talks" means and every screen using it follows, instead of twenty rules that each spell it out and slowly stop agreeing. A named condition can be built from other named conditions, and the page tells you whether each one holds right now.

Use a rule when the screen can work it out: the clock, the data, the session. Use a cue when a person decides — the moment to cut to the keynote, the doors opening early, the thing nobody could have predicted. Rules are how a room runs itself between the moments that need somebody.