# Layouts

Every layout in the kit, picked by a section from what it contains.

A `Section` reads its children and picks one of nine layouts, eight from what it holds and `columns` on request. Each block
declares a kind and a minimum useful width, and a layout collapses when its
own container can no longer give its blocks that width. Resize the window, or
open the sidebar, and each specimen answers to the space it actually has.

## Opener

## Signal survives the quarry wall

A week of logging at the Harrow pit showed that a 433 MHz link holds through
forty metres of basalt when the antenna is kept clear of the face.

_Receiver mast at the north bench, looking back toward the crusher. The logger sits in the grey enclosure at its foot._

## Column

The survey ran for seven days in March, with a transmitter on the pit floor
and three receivers stepped up the benches. Each receiver logged signal
strength and packet loss once a second, which gave a little over 1.8 million
samples by the end of the week.

Two of the three receivers never dropped below the level needed for a clean
decode. The third, set hard against the rock on the middle bench, lost about
one packet in nine until it was moved a metre out from the face.

## Spread

### Antenna clearance

Moving the middle receiver one metre off the face cut its packet loss from
11.2% to 0.4%. Nothing else about the installation changed: same cable, same
height, same radio.

_Packet loss at the middle bench before and after the move, hourly means._

## Plate

_Harrow pit from the eastern rim at first light. Transmitter position is the lit patch on the floor, left of centre._

## Gallery

_North bench receiver, 12 m above the floor._

_Middle bench receiver after it was moved off the face. The longer caption shows that images stay aligned across the row._

_Top bench receiver, 38 m above the floor._

## Stat row

**1.8M** samples logged. one per second, three receivers

**40 m** of basalt between ends

**0.4%** packet loss. middle bench, after the move

**7** days on site

## Margin notes

Signal strength fell by roughly 6 dB for each bench climbed, which is close
to what free-space loss alone predicts for the added distance.

Free-space loss rises 6 dB for each doubling of distance.

The rock itself contributed less than expected. Basalt at Harrow is dry and
dense, and the path the signal takes is mostly over the lip of each bench
rather than through it.

Rain on the fourth day made no measurable difference to any receiver.

11 mm fell between 02:00 and 09:00, measured at the site office gauge.

## Closer

> A metre of air did more for the link than any amount of transmit power.

Logs and the analysis notebook are kept with the survey record.

## Live block

A `Live` block holds the place of an interactive figure and lays out as one.

Drag the receiver up the benches to see predicted signal strength against
the logged values at each position.

**link-budget explorer.** Predicted and logged signal strength by receiver height.

## Drawn figures

A `Figure` or a `Live` block takes a chart, a plate or an interactive as its
child and sets a title above it. The child sets its own height.

Loss at the middle bench fell as soon as the receiver came off the face.

**Packet loss by day.** Daily mean at the middle bench. The receiver was moved on day four.

## Table and panels

A `Table` wraps a Markdown table and sets it on the wide column; `Panels`
holds a row of short definitions. Both sit in a column with the prose around
them.

Each receiver kept its own log for the week.

| Receiver     | Height | Packet loss |
| ------------ | ------ | ----------- |
| North bench  | 12 m   | 0.2%        |
| Middle bench | 24 m   | 0.4%        |
| Top bench    | 38 m   | 0.3%        |

_Receivers at Harrow, March survey._

**Transmitter.** Pit floor, 433 MHz, one packet a second.

**Receiver.** Bench mast with a logger at its foot.

**Link.** Forty metres of basalt between the two.

## Chapter and prose parts

A `Chapter` groups the sections of one part of a long article and sets them
a break apart. Inside it, `Kicker` marks a numbered heading, `Lead` opens the
part, `Formula` holds TeX, `Segue` hands over to the next part and
`Colophon` sets method notes small.

Part II · The benches

### Three receivers

Each bench adds twelve metres of height and a little under 6 dB of loss.

Loss in decibels follows distance.

`L = 20\, \log_{10}(d)`

Height explains most of the loss, and the rock explains the rest.

Signal strength was logged once a second by each receiver and averaged by the hour.

## Rhythm

Inside an `Article` the sections are planned together. The first becomes the
opener and a last section with a pull quote the closer; spreads alternate
sides; a plate that directly follows a plate is set as a column instead.

## Seven days at Harrow

What a week of logging says about running a low-power link out of a working
quarry.

The transmitter went onto the pit floor on a Monday. By Tuesday the first
two receivers were up and logging.

_Pit floor, day one._

The third receiver went in on Wednesday, against the face of the middle
bench, and immediately began dropping packets.

_Middle bench, before the move._

**11.2%** loss against the face

**0.4%** loss one metre out

**28×** improvement

_Eastern rim, day four, after rain._

_A second plate in a row is set as a column._

> Keep the antenna off the rock.

## Columns

A section set with `layout="columns"` puts its prose in balanced columns: two
on the wide column, three in a very wide article, one when the slot cannot
hold two. Headings, figures and quotes span all of them, and the browser
keeps at least two lines of a paragraph at the top or bottom of a column.
It is never picked from contents; a run of long prose asks for it.

### Where the signal went

The receiver on the middle bench lost about one packet in nine for the first
four days. The cable was sound, the radio was the same model as on the other
two benches, and the logger recorded no brown-outs. What differed was the rock:
the mast stood hard against the face, and the first Fresnel zone of the link
crossed it.

Moving the mast one metre out cut the loss to 0.4%. The change was made once,
on the fifth morning, so the experiment has no replication and no control
bench; the other two receivers stayed where they were and stayed clean.

A second explanation is still open. The middle bench is also the one nearest
the crusher, and the crusher ran on four of the seven days. A week with the
crusher idle would separate the two causes, and has not been run.

_Packet loss on the middle bench by hour, before and after the move._

The other two receivers logged between 1.1 and 1.4 million samples each and
never dropped below the level needed for a clean decode. Their antennas stood
more than three metres from any face, which is the clearance the working
rule now asks for.

### What was kept

The raw logs, the notebook that reduced them and the receiver settings are
kept with the survey record. The hourly means in the figure are computed from
the raw logs, not taken from the logger's own summary, which rounds loss to
the nearest whole percent and would hide the change after the move.

## Drop cap

`Prose dropcap` sets the first paragraph with a three-line initial. The letter
is the paragraph's own first letter, styled in place, so screen readers read
the word whole and copying the text copies it whole. A feature piece sets one
on the first paragraph of each chapter by itself.

Receivers fail in ways that read as weather. A link that holds for a week and
then drops for an afternoon is first blamed on rain, then on interference, and
only after both are ruled out on the place the antenna stands.

A second paragraph shows that only the first one takes the initial.

## Margin caption

In a column section with `captions="margin"` (or `captionAt="margin"` on one
figure), the caption sits in the outer margin beside its figure, level with
its foot, when the article is wide enough to have a margin. Narrower, it sits
below. Captioned figures are numbered in order through the piece, and `Ref`
points at one by id:  is the figure below. Numbers
show where a piece asks for them; here, on this page.

The packet loss on the middle bench fell from 11.2% to 0.4% after the mast
moved one metre. The figure below shows the hourly means for both periods,
with the move marked.

_Packet loss on the middle bench, hourly means, before and after the mast was moved one metre out from the face._

Nothing else about the installation changed between the two periods.

## Sidebar

A `Sidebar` is a boxed aside for an explanation that can be skipped: a
definition, a method, a worked example. It is set off by a tone and space, not
a border. Alone with its text in a section it sits beside it (a spread); with
other blocks, between paragraphs.

The link budget decides whether a receiver hears the transmitter at all, and
it is worked out in decibels so that gains and losses can be added.

**A worked link budget**

Transmit power of 14 dBm, plus 3 dBi of antenna gain at each end, minus 98 dB
of path loss over 1.2 km, leaves a received level of -76 dBm.

A receiver that decodes down to -110 dBm then has a margin of 34 dB.

Between paragraphs, the same box spans the text measure.

**Fresnel zone**

The ellipsoid around the straight line between two antennas inside which an
obstacle takes a measurable share of the signal.

A paragraph that follows it.

## Piece cards

`PieceCard` shows a piece from its frontmatter: a chip in a sentence
(), a card, or the lead.

## Sizes and slots

Every block takes `size`: `compact`, `normal` or `expanded`. Left unset, it
picks from the width of its own container; set, it is forced. Compact drops
secondary detail (credits, attributions, later lead paragraphs) and shows a
short form of text where one is given, before it shrinks type. The grids below
show each block in a narrow, a medium and a wide slot, at each size and left
to choose.

### Prose

A paragraph of running text, long enough to wrap in the narrowest slot and to
show where a word may break when the container is small.

### Lead

Does a 433 MHz link hold through forty metres of rock?

A week of logging at the Harrow pit says it does, with the antenna clear of the face.

### Figure

Figure: Looking up through a dense Antarctic beech canopy, branches and moss-covered trunks closing out the sky.

_A photograph opens in a viewer on a click, Enter or Space; Esc closes it._

### Live

_Predicted and logged signal strength by receiver height._

### Stat

**AUD 10–30** Block I. A lead from any detector's headphone jack to a phone. Rough estimate

### PullQuote

> Moving the middle receiver one metre off the face cut its packet loss from 11.2% to 0.4%.

### Note

Measured on the middle bench with the logger at the foot of the mast, one reading a second for seven days.

### Table

| Receiver     | Height | Packet loss |
| ------------ | ------ | ----------- |
| North bench  | 12 m   | 0.2%        |
| Middle bench | 24 m   | 0.4%        |

_Receivers at Harrow, March survey._

### Panels

**Transmitter.** Pit floor, 433 MHz, one packet a second.

Bench mast with a logger at its foot.

### Formula, Segue and Colophon

`L = \frac{d}{v}`

Which of the two readouts fails first when the noise is raised?

Three runs with seeds 1 to 3, 96 cases each, on the same board; the figures are means over the runs.

### Sidebar

**A worked link budget**

Transmit power of 14 dBm, plus 3 dBi of antenna gain at each end, minus 98 dB
of path loss, leaves -76 dBm.

A receiver that decodes down to -110 dBm then has a margin of 34 dB.

### Piece card

### Columns

### A heading across the columns

The first paragraph of a stretch of long prose, long enough to wrap in the
narrowest slot and to be balanced across columns in the widest.

The second paragraph carries on, so the columns have something to balance
against, and a reader can see where a column ends.

### Drop cap

Receivers fail in ways that read as weather, and a link that holds for a week
and then drops for an afternoon is first blamed on rain.

### Margin caption

A paragraph before the figure.

_A caption that sits in the margin when there is one._

## Surface blocks

Four frames for content that is a thing in fact: a listing, a plotted sheet, a
screen, an interface. Each is a block like the others, so it takes `size` and
sits in any slot, and each is for use sparingly.

### Printout

### Plotter

### Console

### Window

Gain 12 dB, squelch off, logging once a second to the enclosure's card.