Room for Drama
Room for DramaOutside, Actually.

San Francisco (demonstration)

The roof terrace with nothing to hide behind

DemonstrationEdit the plan

When the terrace table is open for dinner outside

Each month read on its 30th, or the last day of a shorter month · 2026
Dinner outside, through the year. The terrace table cannot be read for dinner outside in any month yet. 0 months open, 0 shut, 12 unanswered.JanFebMarAprMayJunJulAugSepOctNovDec0months open
  • Works
  • Tight
  • No
  • Unanswered

Rings run from 3 pm at the centre to 11 pm at the edge. Schematic: A shape of the facts, not a picture of a particular date or a measured outline. Each month is its 30th, or the last day of a shorter month, not a promise about the other days.

The terrace table cannot be read for dinner outside in any month yet.

  • All year: cannot be answered, because privacy exposure has not been recorded here.
The same year as a table
MonthOpenUsable hoursWhat decides it
January 30——Privacy exposure closes it
February 28——Privacy exposure closes it
March 30——Privacy exposure closes it
April 30——Privacy exposure closes it
May 30——Privacy exposure closes it
June 30——Privacy exposure closes it
July 30——Privacy exposure closes it
August 30——Privacy exposure closes it
September 30——Privacy exposure closes it
October 30——Privacy exposure closes it
November 30——Privacy exposure closes it
December 30——Privacy exposure closes it

The terrace table for Dinner outside. October 30, 2026 at 1:02 AM, America/Los_Angeles. Now on this date. Confidence low. Cause: What it feels like is what stops it.

Sit down, eat, and still want to be there ninety minutes later.

Dinner outside

We cannot say yet whether The terrace table works for dinner outside, because privacy exposure is still unanswered.

“Unanswered” means missing evidence prevents a reliable verdict. The questions that need answering are listed below.

  • Close to the limit What it feels like is about 55°F — 4°F under (at least 59°F)Estimate: Worked out, but at least one input was assumed rather than measured.

    The air is 56.8°F. It feels about 1.8 degrees cooler than the air, near 55°F.

    Nobody finishes a meal in a seat that is cooking.

  • Close to the limit Wind is about 9 mph — 1 mph over (at most 9 mph)Estimate: Worked out, but at least one input was assumed rather than measured.

    Nothing upwind is doing much; this spot gets close to the open wind speed.

    Wind takes napkins, candles and conversation before it takes comfort.

Still unanswered:Privacy exposureNoise
Why these are unanswered
  • Privacy exposure. No observer has been placed yet. Privacy is a question about who can see in from where, and until one is marked there is nothing to answer.
  • Noise. Nobody has told us what the noise is like here, and it is not something geometry can infer.

What would change this. This stays unanswered until privacy exposure is known. No observer has been placed yet. Privacy is a question about who can see in from where, and until one is marked there is nothing to answer.

Confidence low (38%). Treat this as a starting hypothesis, not an answer.

What this answer was built from
Use
Dinner outside
Place
The terrace table
Weather
Coastal Mediterranean, dry summers, salt air (scenario shape), scenario (from a published source)
Method
Traced geometry and radiation at this hour, measured against what this use needs
Try a changeMake a document

What next here

Testing a change

3 of 8 milestones. One thing is still in the way of privacy, sightlines and screening advice.

  1. Getting an honest answer3 of 4 done
    • Placed on the earth — done. The sun is in the right part of the sky for this address.
    • A spot to read — done. 1 spot is being read.
    • What stands in the way — done. 3 things are drawn, so the shade is worked out rather than assumed.
    • Who can see in — not done. Privacy is the most common reason a terrace goes unused, and it cannot be answered without knowing which windows look at it.
  2. Testing a changeyou are here0 of 2 done
    • A change tested — not done. Testing a change before buying it is the cheapest part of the whole exercise.
    • A forecast checked — not done. A forecast nobody goes back and checks is a guess with a date on it.
  3. Proving it well enough to spend money0 of 2 done
    • Corrected against reality — not done. After a couple of rechecks the model can be corrected against this specific garden.
    • Photographed — not done. A photograph is what turns a drawing somebody sketched into a drawing somebody can check.
  1. Mark the windows that look at this spot

    Being overlooked is one of the most common reasons a good terrace goes unused, and it is the one reading here that cannot be inferred from anything else.

    a tap per windowneeded for privacy, sightlines and screening advice
  2. Get this place to Annotated

    A photographic check on the mapped sightlines and geometry. That makes privacy estimates more credible, but does not replace measured dimensions. Right now: Sightlines drawn from mapped observers remain provisional until their positions and heights are checked. The drawn heights have not been verified against the actual place. What is missing: Photographs of the actual place: None annotated. A photograph is how a drawn height gets checked against a real one. Photographs that know where they were taken from: No photograph has a position and a bearing, so what is marked on them cannot become a measurement. Who can see in: Nobody marked. Privacy is a question about who sees in from where, and the product will not invent one.

    3 things left
  3. Test a change before you buy anything

    The plan is good enough to ask what a screen, a tree or a canopy would actually do here — including the ones that turn out to be a waste of money.

    a few minutes, nothing bought
3 more things worth doing here
  1. Add a second spot and compare the two

    One spot gives a verdict; two give a reason. Most of what is worth learning here is why one corner works at six and the other never does.

    one tap, then read across
  2. See how this spot changes through the year

    A terrace that works in July can be unusable in September for reasons that have nothing to do with temperature. The same geometry read in four seasons is the cheapest insight here.

    a read, nothing to enter
  3. Photograph the spot and mark what is in shot

    A photograph is what separates a drawing you sketched from a drawing somebody else can check. It is also how heights get corrected without a ladder.

    one photograph

How well this place is described

Mapped
  1. Quick
  2. Mapped
  3. Annotated
  4. Measured

Shade, sky view and daily light are computed from the geometry on this plan. Unmeasured dimensions remain estimates, not verified measurements.

Sightlines drawn from mapped observers remain provisional until their positions and heights are checked. The drawn heights have not been verified against the actual place.

To reach Annotated

  • Photographs of the actual place: None annotated. A photograph is how a drawn height gets checked against a real one.
  • Photographs that know where they were taken from: No photograph has a position and a bearing, so what is marked on them cannot become a measurement.
  • Who can see in: Nobody marked. Privacy is a question about who sees in from where, and the product will not invent one.

What that buys: A photographic check on the mapped sightlines and geometry. That makes privacy estimates more credible, but does not replace measured dimensions.

Everything that was checked
  • Quick A place on the earth: San Francisco (demonstration), with a time zone, which is what makes an hour an hour.
  • Quick Somewhere to read: 1 spot marked. A reading is a height at a place, and moving it moves the answer.
  • Mapped The things in the way: 3 objects drawn. The house counts, and so does the neighbour's.
  • Mapped What is underfoot: 1 of 1 readings sit on a mapped surface.
  • Mapped Where the property ends: Drawn, so the product can tell you when the cheapest fix stands on someone else's land.
  • Annotated Photographs of the actual place: None annotated. A photograph is how a drawn height gets checked against a real one.
  • Annotated Photographs that know where they were taken from: No photograph has a position and a bearing, so what is marked on them cannot become a measurement.
  • Annotated Who can see in: Nobody marked. Privacy is a question about who sees in from where, and the product will not invent one.
  • Measured Heights off a tape: 2 of 3 objects are measured. The remaining 1 are this product's own inference, so no figure here is a survey figure yet.
  • Measured Something checked against the day: Nothing has been checked against a real afternoon. The model is unopposed, which is not the same as being right.

What to measure next

Ranked by how far the answer moves, not by a checklist
  1. 1Who can actually see inSit in the chair

    Sit where you actually sit and look around. Mark the one window or path you can see a person from. Until one exists, privacy is a question we are not answering rather than one that is fine.

    This is not a refinement. Nothing can answer this question until somebody marks it.

    Mark it on the plan

    Currently assuming: nobody is overlooking this. Tested across one confirmed observer changes the whole privacy answer.

Drawn
Worked out
Terrace paversThe bayParapet3 ft 7 inWest parapet3 ft 7 inStair head8 ft 6 inThe terrace table9121518N16 ft
BuildingNot yours (dashed edge)CanopySun’s path today (schematic)Can see inProposed

To scale · to the measurements given: Drawn to scale from the measurements given. Not a survey. It is not a survey.

Reading at

1:02 AM·October

The sun is down.

October 30 · The terrace table Schematic: A shape of the facts, not a picture of a particular date or a measured outline.
  • Unanswered (hatched)

The sun reaches the terrace table at 7:34 AM. The terrace table loses the sun at 6:11 PM.

The evidence behind this whole day is thin, so nothing on the strip is marked out as worse than the rest. The shape of the day is still the model’s best reading; the confidence in it is low, and it stays low until something it needs is measured.

The terrace table cannot be answered for dinner outside yet.

Nothing in October clears what dinner outside needs here.

The same day as a table
The same day hour by hour, in text.
HourSunVerdictConfidenceSun position
12:00 AMdarkUnanswereda starting hypothesis
3:00 AMdarkUnanswereda starting hypothesis
6:00 AMdarkUnanswereda starting hypothesis
8:00 AMin sunUnanswereda starting hypothesis4° up, bearing 111°
9:00 AMin sunUnanswereda starting hypothesis15° up, bearing 121°
10:00 AMin sunUnanswereda starting hypothesis24° up, bearing 133°
11:00 AMin sunUnanswereda starting hypothesis32° up, bearing 147°
12:00 PMin sunUnanswereda starting hypothesis37° up, bearing 164°
1:00 PMin sunUnanswereda starting hypothesis38° up, bearing 182°
2:00 PMin sunUnanswereda starting hypothesis36° up, bearing 200°
3:00 PMin sunUnanswereda starting hypothesis30° up, bearing 216°
4:00 PMin sunUnanswereda starting hypothesis22° up, bearing 230°
5:00 PMin sunUnanswereda starting hypothesis12° up, bearing 241°
6:00 PMin sunUnanswereda starting hypothesis2° up, bearing 251°
9:00 PMdarkUnanswereda starting hypothesis
When the sun arrives and leaves, to the minute.
WhenWhat happens
7:34 AMThe sun reaches the terrace table at 7:34 AM.
6:11 PMThe terrace table loses the sun at 6:11 PM.

Direct sun at the terrace table, month by month

Schematic · computed: Computed from this place's own plan and climate, then drawn as a shape rather than to scale.

The terrace table gets the most direct sun in May, about 15 hours on the 30th, or the last day of a shorter month, and the least in November, about 9 hours.

  • Sun reaches the spot (solid)
  • Sun up, something in the way (lined)
  • Sun down (empty)
The grid, in a table
Hours of direct sun at The terrace table on the 30th, or the last day of a shorter month, and when.
MonthDirect sunWhen
January10 h8:00 AM–6:00 PM
February12 h7:00 AM–7:00 PM
March13 h7:00 AM–8:00 PM
April13 h7:00 AM–8:00 PM
May15 h6:00 AM–9:00 PM
June15 h6:00 AM–9:00 PM
July14 h7:00 AM–9:00 PM
August13 h7:00 AM–8:00 PM
September11 h8:00 AM–7:00 PM
October11 h8:00 AM–7:00 PM
November9 h8:00 AM–5:00 PM
December9 h8:00 AM–5:00 PM
The sun’s path on the plan, October 30, in a table
Where the sun is at each hour it is up.
HourComes fromHeight
8:00 AME (111°)4° up
9:00 AMSE (121°)15° up
10:00 AMSE (133°)24° up
11:00 AMSE (147°)32° up
12:00 PMS (164°)37° up
1:00 PMS (182°)38° up
2:00 PMS (200°)36° up
3:00 PMSW (216°)30° up
4:00 PMSW (230°)22° up
5:00 PMSW (241°)12° up
6:00 PMW (251°)2° up

What the place is doing at 1:02am in October

Coastal Mediterranean, dry summers, salt air (scenario shape).

What it feels like

Estimate: Worked out, but at least one input was assumed rather than measured.cool

55°F

Felt temperature

The bar is the 0–100 reporting scale, mapped from the reading above: 0 is none of it, 100 the top of the scale.

The air is 56.8°F. It feels about 1.8 degrees cooler than the air, near 55°F.

What is doing it

What makes this reading 55°FA running total. It starts at 52°F with every named influence removed, then each influence is added in turn.52°Fshaded air+3°FAir moveme…55°Fwhat it is

Take every named influence away and this reads 52°F. Air movement adds 3°F. That lands at 55°F.

How this was worked out
Air
56.8°F from Coastal Mediterranean, dry summers, salt air (scenario shape), scenario (from a published source)
Radiant
44.3°F mean radiant temperature
Method
Six-directional radiation integration, then ASHRAE operative temperature (from a published source)

Direct sun

Worked out: Computed from the geometry, the sun and the weather you gave it.after dark

0 W/m²

Beam reaching this point

The bar is the 0–100 reporting scale, mapped from the reading above: 0 is none of it, 100 the top of the scale.

The sun is down.

How this was worked out
Sun
below the horizon
Traced
0% of the beam reaches this point

Glare

Worked out: Computed from the geometry, the sun and the weather you gave it.none

0

On the 0–100 reporting scale: 0 is none of it, 100 the top of the scale. Not a measurement.

The bar is the 0–100 reporting scale, mapped from the reading above; the tick is where this use stops working.

Nothing bright is in the line of sight at this hour.

How this was worked out
Facing
260deg

Privacy exposure

Unanswered: Something this depends on has not been given, so nothing is being guessed.

No observer has been placed yet. Privacy is a question about who can see in from where, and until one is marked there is nothing to answer.

Wind

Estimate: Worked out, but at least one input was assumed rather than measured.gusty

9 mph

Wind at the seat

The bar is the 0–100 reporting scale, mapped from the reading above: 0 is none of it, 100 the top of the scale.

Nothing upwind is doing much; this spot gets close to the open wind speed.

How this was worked out
Reference
8.4 mph at 10 m from Coastal Mediterranean, dry summers, salt air (scenario shape), scenario (from a published source)
At seat height
9.5 mph before shelter
Shelter
0% removed by what stands upwind

View

Worked out: Computed from the geometry, the sun and the weather you gave it.open

100

On the 0–100 reporting scale: 0 is none of it, 100 the top of the scale. Not a measurement.

The bar is the 0–100 reporting scale, mapped from the reading above: 0 is none of it, 100 the top of the scale.

The view from this seat is essentially intact.

What is doing it

  • The bay

    +100
How this was worked out
Corridors
The bay
Method
Rays sampled across each corridor's angular window

Surface heat

Estimate: Worked out, but at least one input was assumed rather than measured.comfortable

57°F

Surface temperature

The bar is the 0–100 reporting scale, mapped from the reading above: 0 is none of it, 100 the top of the scale.

The surface is running about 57°F.

How this was worked out
Surface
Terrace pavers
Method
Absorbed shortwave against air temperature, flattened by air movement (from a published source)

Wetness

Estimate: Worked out, but at least one input was assumed rather than measured.dries-fast

1 × the open

Drying time against the open

The bar is the 0–100 reporting scale, mapped from the reading above; the tick is where this use stops working.

Sun and air movement here are enough to dry the surface at an ordinary rate.

How this was worked out
Drying rate
0.000 in/hour here
Method
Penman combination equation on the radiation and air actually reaching this spot (from a published source)

Light after dark

Unanswered: Something this depends on has not been given, so nothing is being guessed.

No exterior or window light has been placed yet, so what changes after dark is unanswered.

Ask something else about this place

Everything on the plan, in text
ThingReadingStandingNote
Parapet3 ft 7 inMeasured: Taken with a tape, a compass or a clock at the place.100 per cent solid.
West parapet3 ft 7 inMeasured: Taken with a tape, a compass or a clock at the place.100 per cent solid.
Stair head8 ft 6 inYou said so: You told us this, and it beats what the model worked out.100 per cent solid.