# Docs Markup and Syntax Guide¶

The ODK documentation is built using Sphinx, a static-site generator designed to create structured, semantic, and internally consistent documentation. Source documents are written in reStructuredText, a semantic, extensible markup syntax similar to Markdown.

Note

Sphinx and reStructuredText can be very flexible. For the sake of consistency and maintainability, this guide is highly opinionated about how documentation source files are organized and marked up.

## Indentation¶

Indentation is meaningful in Sphinx and reStructuredText text. Usually, indenting a section means that is "belongs to" the line it is indented under. For example:

.. figure:: path-to-image.*

This is the caption of the figure. Notice that it is indented under the line defining the figure.


The rules for indentation are:

• Use spaces, not tabs.
• Generally, indent two spaces.

The exception to the two spaces rule is Ordered (numbered) lists, where indentation follows the content of the list item.

1. This is a list item.

This is some additional content related to Item 1. Notice that it is indented to the same column as the first line of content. In this case, that's three (3) spaces.

.
.
.

10. The tenth item in a list.

This related content will be indented four spaces.


## Documentation Files¶

Sphinx document files have the .rst extension. File names should be all lowercase and use hyphens (not underscores or spaces) as word separators.

Normally, the title of the page should be the first line of the file, underlined with equal-signs.

Title of Page
================

Page content is here...


You can alternatively wrap the title in two lines of asterisks, in some cases. (This should not be your default choice.)

*******************
Title of Page
*******************

Page content here.


The asterisks style is useful when you are combining several existing documents and don't want to change every subsection headline. Or, you can use it when you are working on a document that you have reason to think might be split into separate documents in the future.

Important

If you use the double-asterisks style, your major section headlines (<h2>) should use the equal-signs underline style. This allows major sections to be easily promoted to individual pages.

See Sections and Titles for more details.

## Tables of Content¶

The toctree directive defines a table of content. The content of a toctree is a list of page file names, without the .rst extension. When rendered, the toctree becomes an unordered list of page links, including links to sections and subsections of the included pages.

.. toctree::

page-name
another-page
this-other-page


The depth of section and subsection links to display in the output can be controlled using the maxdepth attribute. We typically use a depth of 2, but you should use your judgment if you feel it should be more or less in any given context.

.. toctree::
:maxdepth: 2

this-page
that-page
thick-page
flat-page


The TOC Tree

The Sphinx documentation includes information about a number of other toctree attributes.

### Secondary tables of content¶

Collections of documents are sometimes given their own table of content on an individual page. (See, for example, Setting Up ODK Collect and Using ODK Collect.)

In these cases, the page containing the toctree serves as a sort of intro page for the collection. That intro must, itself, be included in the Sidebar navigation menu.

The contents of a toctree appear as section links in another toctree it is included in. That is, if a toctree in index.rst lists collect-using, and collect-using.rst has a toctree, then the contents of that second toctree will appear in the Sidebar navigation menu, as sub-items to Using ODK Collect. (Indeed, this is precisely the case in the docs currently.)

### How ODK Docs uses main and secondary tables of content¶

• Major topics get a toctree in index.rst

Major topics include things like:

• Each major product (Collect, Aggregate, Briefcase)
• Large, general categories like Contributing

Major topic tables of content include both sub-collection intro pages and also individual pages that don't fit into a sub-collection.

The caption attribute of the toctree directive defines the section label in the Sidebar navigation menu.

• Within a large topic, documents are grouped into collections of related pages, defined by a toctree on a topic intro page.

Intro pages (pages that contain secondary toctree directives) may include additional content, introducing the collection or providing contextual wayfinding. However, this is not always necessary or desirable. Use your judgment, and avoid stating things just for the sake of having some text. ("Here are the pages in this collection.")

We also (very occasionally) include toctree directives in sub-collection pages.

Tip

If it not obvious where a new document should appear in the navigation, the best practice is to simply ask about it in the GitHub issue driving the new page.

Note

For wayfinding purposes, we sometimes create an Unordered (bullet) lists of page links rather than a toctree directive. (For example, see collect-intro. We do this when using a toctree would create redundant links in the Sidebar navigation menu.

Why are the docs files not grouped into folders in the source?

We use toctree directives as our primary way of organizing the documentation for readers. We do not organize the source rst files into subfolders.

The reason is that if we put them into topic-related subfolders, it would affect the URI of the document. Keeping all of our document files in a single flat directory results in a flat URI structure. Every page's URI looks like docs.opendatakit.org/page-name.

If we used subdirectories, then our URIs would look like docs.opendatakit.org/subdirectory-name/page-name. This would mean that our URIs would change every time we moved a document from one folder to another, greatly increasing the time cost and broken-link risk of reorganizing the docs.

## Sections and Titles¶

• the text of the headline, followed by
• a line filled with a single character.

Each level in a headline hierarchy uses a different character:

Title of the Page - <h1> - Equal Signs
=========================================

Major Section - <h2> - Hyphens
---------------------------------

Subsection - <h3> - Tildes
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Sub-subsection - <h4> - Double Quotes
"""""""""""""""""""""""""""""""""""""""

Sub-sub-subsection - <h5> - Single Quotes
''''''''''''''''''''''''''''''''''''''''''''


If you need to combine several existing pages together, or want to start a single-page doc that you think might be split into individual pages later on, you can add a top-level title, demoting the other headline types by one.

************************************************
Page Title - <h1> - Asterisks above and below
************************************************

Major Section - <h2> - Equal Signs
=======================================

Subsection - <h3> - Hyphens
---------------------------------

Sub-subsection - <h4> - Tildes
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Sub-sub-subsection - <h5> - Double Quotes
"""""""""""""""""""""""""""""""""""""""""""""

Sub-sub-sub-subsection - <h6> - Single Quotes
''''''''''''''''''''''''''''''''''''''''''''''''''


In either case, the underline of characters needs to be longer than the line of text. In the case of the asterisks, the two lines of asterisks need to be the same length.

Note

The exact order of underline characters is flexible in reStructuredText. However, this specific ordering should be used throughout the ODK documentation.

### Section labels¶

In order to facilitate efficient Cross referencing, sections should be labeled. This is done on the line above the section title. The format is:

• two dots
• underscore
• section label
• lowercase
• hyphen separators
• a single colon
.. _section-label:

Section Title
----------------

Lorem ipsum content of section blah blah.


The section label is usually a slugified version of the section title.

Section titles must be unique throughout the entire documentation set. Therefore, if you write a common title that might appear in more than one document (Learn More or Getting Started, for example), you'll need to include additional words to make the label unique. The best way to do this is to add a meaningful work from the document title.

ODK Aggregate
===============

ODK Aggregate is a server application...

.. _aggregate-getting-started:

Getting Started
-----------------


## Basic Markup¶

### Escaping characters¶

Markup characters can be escaped using the \ character.

*Italic.*

\*Not italic, surrounded by asterisks.\*


Italic.

*Not italic, surrounded by asterisks.*

### Emphasis and Inline Literal¶

Single asterisks for *italic text* (<em>).

Double asterisks for **bold text** (<strong>).

Double back-ticks for inline literal text (<code>).


Single asterisks for italic text ( <em> ).

Double asterisks for bold text ( <strong> ).

Double back-ticks for inline literal text ( <code> ).

Note

The bold, italic, and inline literal styles do not carry semantic meaning. They should not be used when a more semantically appropriate markup construct is available; for example, when writing about GUI text.

### Lists¶

#### Unordered (bullet) lists¶

Bulleted lists ( <ul> ):

- use hyphens
- are unindented at the first level
- must have a blank line before and after

- the blank line requirement means that nested list items will have a blank line before and after as well

- you may *optionally* put a blank line *between* list items


Bulleted lists ( <ul> ):

• use hyphens
• are unindented at the first level
• must have a blank line before and after
• the blank line requirement means that nested list items will have a blank line before and after as well
• you may optionally put a blank line between list items

#### Ordered (numbered) lists¶

Numbered lists ( <ol> ):

1. Start each line with a number and period
2. Can begin on any number
3. Must have a blank line before and after
4. Can have nested sub-lists

a. nested lists are numbered separately
b. nested lists need a blank line before and after

#. Can have automatic number with the # character.


Numbered lists ( <ol> ):

1. Start each line with a number and period
2. Can begin on any number
3. Must have a blank line before and after
4. Can have nested sub-lists
1. nested lists are numbered separately
2. nested lists need a blank line before and after
5. Can have an automatic number with the # character.

Note

See Ordered and unordered lists in the Style Guide for details on when to use ordered and unordered lists.

#### Definition Lists¶

Definition list ( <dl> )
a list with several term-definition pairs

Terms
should not be indented

Definitions
should be indented under the term

Line spacing
there should be a blank line between term-definition pairs

Definition list ( <dl> )
a list with several term-definition pairs
Terms
should not be indented
Definitions
should be indented under the term
Line spacing
there should be a blank line between term-definition pairs

### Paragraph-level Markup¶

Paragraphs are separated by blank lines. Line breaks in the source code do not create line breaks in the output.

This means that you *could*, in theory,
include a lot of arbitrary line breaks
These line breaks would not appear in the output.
Some people like to do this because they have been trained
to not exceed 80 column lines, and they like
to write .txt files this way.

There is **no reason** to put a limit on line length in source files for documentation, since this is prose and not code. Therefore, please do not put arbitrary line breaks in your files.


Paragraphs are separated by blank lines. Line breaks in the source code do not create line breaks in the output.

This means that you could, in theory, include a lot of arbitrary line breaks in your source document files. These line breaks would not appear in the output. Some people like to do this because they have been trained to not exceed 80 column lines, and they like to write .txt files this way. Please do not do this.

There is no reason to put a limit on line length in source files for documentation, since this is prose and not code. Therefore, please do not put arbitrary line breaks in your files.

#### Block Quotes¶

This is not a block quote. Block quotes are indented, and otherwise unadorned.

This is a block quote.


This is not a block quote. Block quotes are indented, and otherwise unadorned.

This is a block quote. — Adam Michael Wood

#### Line Blocks¶

| Line blocks are useful for addresses,
|
| Each new line begins with a
| vertical bar ("|").
|     Line breaks and initial indents
|     are preserved.

Line blocks are useful for addresses,

Each new line begins with a
vertical bar ("|").
Line breaks and initial indents
are preserved.

#### Tables¶

##### Grid style¶
+------------+------------+-----------+
+============+============+===========+
| body row 1 | column 2   | column 3  |
+------------+------------+-----------+
| body row 2 | Cells may span columns.|
+------------+------------+-----------+
| body row 3 | Cells may  | - Cells   |
+------------+ span rows. | - contain |
| body row 4 |            | - blocks. |
+------------+------------+-----------+

body row 1 column 2 column 3
body row 2 Cells may span columns.
body row 3 Cells may span rows.
• Cells
• contain
• blocks.
body row 4
##### Simple style¶
=====  =====  ======
Inputs     Output
------------  ------
A      B    A or B
=====  =====  ======
False  False  False
True   False  True
False  True   True
True   True   True
=====  =====  ======

Inputs Output
A B A or B
False False False
True False True
False True True
True True True
##### CSV Table¶

The csv-table role is used to create a table from CSV (comma-separated values) data. CSV is a common data format generated by spreadsheet applications and commercial databases. The data may be internal (an integral part of the document) or external (a separate file).

.. csv-table:: Example Table
:widths: 15, 10, 30

"Albatross", 2.99, "On a stick!"
"Crunchy Frog", 1.49, "If we took the bones out, it wouldn't be
crunchy, now would it?"
"Gannet Ripple", 1.99, "On a stick!"

Example Table
Treat Quantity Description
Albatross 2.99 On a stick!
Crunchy Frog 1.49 If we took the bones out, it wouldn't be crunchy, now would it?
Gannet Ripple 1.99 On a stick!

Some of the options recognized are:

:widths:

Contains a comma or space-separated list of relative column widths. The default is equal-width columns.

The special value auto may be used by writers to decide whether to delegate the determination of column widths to the backend.

In most cases, the best result is either the default or auto. If you're unsure, try it both ways and see which looks better to you.

:header:

Contains column titles. It must use the same CSV format as the main CSV data.

:delim:

Contains a one character string used to separate fields. Default value is comma. It must be a single character or Unicode code.

The only reason to use something other than a comma is when copying large blocks of content from another source that uses a different style. If you are creating new table content yourself, use the comma.

.. csv-table:: Table using # as delimiter
:widths: auto
:delim: #

"Peter"#"A"
"Paul"#"B"

.. csv-table:: Table using | as delimiter
:widths: auto
:delim: |

"Peter"|"A"
"Paul"|"B"

:align:

It specifies the horizontal alignment of the table. It can be left, right or center.

.. csv-table:: Table aligned to right
:align: right

"Peter", "A"
"Paul", "B"

Table aligned to right
Peter A
Paul B
:file:

Contains the local file system path to a CSV data file.

:url:

Contains an Internet URL reference to a CSV data file.

Note

• There is no support for checking that the number of columns in each row is the same. However, this directive supports CSV generators that do not insert "empty" entries at the end of short rows, by automatically adding empty entries.

.. csv-table:: Table with different number of columns in each row

"Peter"
"Paul", "B"

Table with different number of columns in each row
Peter
Paul B
• Whitespace delimiters are supported only for external CSV files.

For more details, refer to this guide on CSV Tables.

Note

In almost all cases, csv-table is the easiest and most maintainable way to insert a table into a document. It should be preferred unless there is a compelling reason to use one of the other styles.

## Sphinx-specific Markup¶

### Roles and directives¶

A role is an inline markup construct that wraps some text, similar to an HTML or XML tag. They look like this:

:rolename:some text


A directive is a block-level markup construct. They look like this:

.. directivename:: additional info or options here
:option: optional-value
:option: optional-value

Content of block here, indented.

This is no longer part of the block controlled by the directive.


Most of the Sphinx-specific and ODK-specific markup will use one or both of these constructs.

### Cross referencing¶

Cross referencing is linking internally, from one place in the documentation to another. This is not done using the Hyperlinks syntax, but with one of the several roles:

:role:target
becomes...
<a href="target">reference title</a>

:role:anchor text <target>
becomes...
<a href="target">anchor text</a>

:doc:
• target is the file name, without the .rst extension
• title is the first headline ( <h1> ) of the page
:ref:
:term:
• Links to items in the Glossary
• target is the term, in the glossary
• title is the term itself

To recap: If you do not include an explicit target, the text inside the role will be understood as the target, and the anchor text for the link in the output will be the title of the target.

For example:

- Link to this document:

- :doc:contributing
- :doc:anchor text <contributing>

- :ref:cross-referencing
- :ref:anchor text <cross-referencing>

- :term:participant
- :term:anchor text <participant>


Several roles are used when describing user interactions.

:guilabel:

Marks up actual UI text of form labels or buttons.

Press the :guilabel:Submit button.

:menuselection:

Marks up the actual UI text of a navigation menu or form select element.

Select :menuselection:Help from menu.


When writing about multi-level menus, use a single :menuselection: role, and separate menu choices with -->.

To save your file, go to :menuselection:File --> Save in the Main Menu.


Note

In some situations you might not be clear about which option (menuselection or guilabel) to use. GUIs in real life can sometimes be ambiguous. The general rule is:

These both render the same on output, so don't worry too much if you get it wrong. Just use your judgment and take your best guess.

:kbd:

Marks up a sequence of literal keyboard strokes.

To stop the local server, type :kbd:CTRL C.

:command:

Marks up a terminal command.

To build the documentation, use :command:sphinx-build.

:option:

Marks up a terminal command option.

The :option:-b html option specifies the HTML builder.

:gesture:

Describes a touch screen gesture.

:gesture:Swipe Left


We have added several custom text roles for writing about forms and the XForms and XLSForm formats.

:th:

Used to refer to a table header cell.

:tc:

Used to refer to a table cell.

External App String Widget
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
The external app widget is displayed when the :th:appearance attribute begins with :tc:ex:.

:formstate:

Specifies the state of the form in ODK Collect, which could be one of the following:

• Blank
• Finalized
• Saved
• Sent
• Deleted
:formstate:Sent


### Other Semantic Markup¶

:abbr:

Marks up an abbreviation. If the role content contains a parenthesized explanation, it will be treated specially: it will be shown in a tool-tip in HTML.

:abbr:ODK (Open Data Kit)

:dfn:

Marks the defining instance of a term outside the glossary.

:dfn:Open Data Kit (ODK) is a suite of open source applications that help organizations engaged in enumerator-mediated data collection.

:file:

Marks the name of a file or directory. Within the contents, you can use curly braces to indicate a "variable" part.

is installed in :file:/usr/lib/python2.{x}/site-packages


In the built documentation, the x will be displayed differently to indicate that it is variable.

:program:

Marks the name of an executable program.

launch the :program:ODK Aggregate Installer


### Images and Figures¶

#### Where to put image files¶

Image files should be put in the /src/img/ directory in the source, and they should be in a subdirectory with the same name as the document in which they appear. (That is, the filename without the .rst extension.)

#### Image compression¶

You must perform lossless compression on the source images. Use one of the following tools to optimize the images:

• ImageOptim is a tool that allows us to optimize the images. It is not format specific which means it can optimize both jpeg as well as png images. You can download it from here . After launching ImageOptim.app, dragging and dropping images into its window gives you an in-place optimized file.
• Pngout is another option for optimizing png images. Installation and usage instructions can be found here .
• Mozjpeg can be used to optimize jpeg images. Installation and related information can be found on this link .

#### Inserting images in a document¶

To place an image in a document, use the image directive.

.. image:: /img/{document-subdirectory}/{file}.*
:alt: Alt text. Every image should have descriptive alt text.


Note the literal asterisk (*) at the end, in place of a file extension. Use the asterisk, and omit the file extension.

#### Inserting images with captions (figures)¶

Use figure to markup an image with a caption.

.. figure:: /img/{document-subdirectory}/{file}.*
:alt: Alt text. Every image should have descriptive alt text.

The rest of the indented content will be the caption. This can be a short sentence or several paragraphs. Captions can contain any other rst markup.


#### Inline images¶

To information on creating inline images, see Substitutions.

#### Image File Names¶

Image file names should:

• be short yet descriptive
• contain only lower case characters and (in Sequentially numbered images only) numbers
• have no spaces
• use hyphens as the separator

Good image file names:

• collect-home-screen.png
• build-data-export-menu.png

• Collect home screen.png
• collect_home_screen.png
• 3987948p2983768ohl84692p094.jpg-large
##### Sequentially numbered images¶

In the case of sequentially numbered images, the numbers should:

• be zero-indexed
• have two digits with leading zeroes
• be separated from the rest of the file name with a hyphen
• be placed at the end of the file name

Good sequentially numbered image file names:

• map-widget-00, map-widget-01, map-widget-02

Bad sequentially numbered image file names:

• 1-map-widget, 2-map-widget
• map-widget_00, map-widget_01
• map-widget-1, map-widget-2

#### Screenshots from ODK Collect¶

If you have set up local Android Tools, you can connect your Android device to your computer and take screenshots from the command line.

• Connect your device via USB
• Enable Developer Settings
• Tap Build number seven (7) times
• Turn on USB Debugging
• Settings ‣ Developer options ‣ USB debugging

Now, at the command line, from the root directory of the odk-docs repo:

python ss.py {document-name}/{image-name}

• {document-name} is the filename (without extension) where the image will be used.
• {image-name} is the name (without extension) given to the image. - follow the Image File Names guidelines

Warning

Make sure you do not overwrite an existing image.

Tip

If you have a problem running ss.py, check to make sure your Python 3 virtual environment is activated.

Tip

Be sure to obscure any personally-identifiable information from screen shots. Crop to the smallest relevant screen area. Annotate screen shots with arrows or circles to indicate relevant information.

### Videos¶

Video files should be put in the /src/vid/ directory in the source, and they should be in a subdirectory with the same name as the document in which they appear. (That is, the filename without the .rst extension.)

The purpose of on page videos is to illustrate complicated user interactions that might be difficult to describe otherwise. Longer tutorial videos should be hosted elsewhere and, if appropriate, linked to from the docs. Therefore:

• The length of the videos must be less than a minute.
• Videos should have no audio.

To insert a video, use the custom video directive.

.. video:: path/to/video

Specify the source path of the video and a descriptive alt content in the video directive. Alternate content is displayed when the video cannot be played. It can contain long texts as well as any other rst content.

.. video:: /vid/{document-subdirectory}/{file}.ext

Alt content. Every video should have descriptive alt content.


The following optional attributes are supported:

:autoplay:

Specifies whether the video should start playing as soon as it is ready. Can take boolean value: true, false, yes or no. Default is no.

It is almost never a good idea to turn autoplay on.

:controls:

Specifies whether the video controls should be displayed. Can take boolean value: true, false, yes or no. Default is yes.

:muted:

Specifies whether the audio output of the video should be muted. Can take boolean value: true, false, yes or no. Default is yes.

:loop:

Specifies whether the video should start over again, every time it is finished. Can take boolean value: true, false, yes or no. Default is no.

:preload:

Specifies if and how the author thinks the video should be loaded when the page loads. Can take one of the following three values: auto, metadata or none.

:poster:

Contains the source address for an image to be shown while the video is downloading, or until the user hits the play button.

Note

Images to be used as poster for a video should be in the same directory as the video and should have a filename like [same-file-name-as-video]-poster.ext.

:class:

Specifies a class for the video element.

For more details on these attributes, see this guide.

To add a video in a document with the above options, you can do the following:

.. video:: /vid/{document-subdirectory}/{file}.ext
:autoplay: yes/no
:controls: yes/no
:muted: yes/no
:loop: yes/no
:class: class-name
:poster:: /vid/{document-subdirectory}/{file}.ext

Alt content. Every video should have descriptive alt content.


#### Capturing video from Collect¶

Android Debug Bridge (ADB) can be used to capture a screen recording from ODK Collect.

$adb shell screenrecord /sdcard/example.mp4  On pressing the enter key the video recording starts. Recording stops automatically after 3 minutes. Since videos have to be less than a minute, press CTRL C to stop the recording. The video file is saved in your Android device to a file at /sdcard/example.mp4 file. To pull the video locally: $ adb pull /sdcard/example.mp4 local/path/to/save/to


Downloadable files should be put in the /src/downloads/ directory in the source, and they should be in a subdirectory with the same name as the document in which they appear. (That is, the filename without the .rst extension.)

To place a downloadable file in a document, use the download role.

See this :download:example script </downloads/contributing/example_script.py> to understand the procedure better.


### Code Samples¶

Use the code-block directive to insert code samples. Specify the language on the same line as the directive for syntax highlighting.

.. code-block:: rst

Use the code-block directive to markup code samples.

.. code-block:: python

print("Hello ODK!")

.. code-block:: console

\$ python --version

.. code-block:: java

public class HelloWorld {

public static void main(String[] args) {
// Prints "Hello, World" to the terminal window.
System.out.println("Hello, World");
}

}


Note

rst code-blocks wrap overflow lines by default. To unwrap overflow lines, use unwrap class with rst code-blocks.

.. code-block:: rst
:class: unwrap


Code-blocks for other languages don't wrap overflow lines. Instead of wrapping, you need to scroll side-ways. To wrap overflow lines with other code-blocks, use wrap class with them.

.. code-block:: python
:class: wrap


### Substitutions¶

Substitutions are a useful way to define a value which is needed in many places.

Substitution definitions are indicated by an explicit markup start (".. ") followed by a vertical bar, the substitution text (which gets substituted), another vertical bar, whitespace, and the definition block.

A substitution definition block may contain inline-compatible directives such as image or replace. For more information, refer this guide.

You can define the value once like this:

.. |RST| replace:: reStructuredText


and then reuse it like this:

We use |RST| to write documentation source files.


Here, |RST| will be replaced by reStructuredText

You can also create a reference with styled text:

.. |slack| replace:: **ODK Slack**
.. slack: https://opendatakit.slack.com


You can use the hyperlink reference by appending a "_" at the end of the vertical bars, for example:

You can ask about your problem in |slack|_.


The rst_epilog in conf.py contains a list of global substitutions that can be used from any file. The list is given below:

• If you want to create a hyperlink reference for ODK Slack, you can use |odk-slack|_.

You can use |odk-slack|_ to ask your questions.


• To create a hyperlink reference for docs related issues, use |docs-issue|_.

If you find a problem, file an |docs-issue|_.


If you find a problem, file an issue.

• To create a hyperlink reference for ODK Forum, use |forum|_.

You can ask support questions in |forum|_.


You can ask support questions in ODK Forum.

• To create a hyperlink reference for contributors guide, use |contrib-guide|_.

Be sure to read the |contrib-guide|_.


Be sure to read the contributors guide.

You can add inline images in the document using substitutions. The following block of code substitutes arrow in the text with the image specified.

The |arrow| icon opens the jump menu.

.. |arrow| image:: /img/{document-subdirectory}/{file}.*
:alt: Alt text.