diff --git a/indoc/Chapter 1/Basics.w b/indoc/Chapter 1/Basics.w
index be170391b..0ea1a1abd 100755
--- a/indoc/Chapter 1/Basics.w	
+++ b/indoc/Chapter 1/Basics.w	
@@ -30,7 +30,7 @@ void Basics::end(void) {
 @h Setting up the memory manager.
 We need to itemise the structures we'll want to allocate:
 
-@e indoc_instructions_MT
+@e settings_block_MT
 @e volume_MT
 @e chapter_MT
 @e section_MT
@@ -49,7 +49,7 @@ We need to itemise the structures we'll want to allocate:
 @ And then expand:
 
 =
-ALLOCATE_INDIVIDUALLY(indoc_instructions)
+ALLOCATE_INDIVIDUALLY(settings_block)
 ALLOCATE_INDIVIDUALLY(volume)
 ALLOCATE_INDIVIDUALLY(chapter)
 ALLOCATE_INDIVIDUALLY(section)
diff --git a/indoc/Chapter 1/Configuration.w b/indoc/Chapter 1/Configuration.w
index 6ef378885..34967e51d 100644
--- a/indoc/Chapter 1/Configuration.w	
+++ b/indoc/Chapter 1/Configuration.w	
@@ -14,7 +14,7 @@ void Configuration::add_instructions_file(filename *F) {
 	ADD_TO_LINKED_LIST(F, filename, instructions_files);
 }
 
-void Configuration::read_instructions(text_stream *target, indoc_instructions *settings) {
+void Configuration::read_instructions(text_stream *target, settings_block *settings) {
 	Instructions::read_instructions(target, instructions_files, settings);
 }
 
@@ -23,18 +23,16 @@ void Configuration::read_instructions(text_stream *target, indoc_instructions *s
 =
 typedef struct cl_state {
 	struct text_stream *target_chosen;
-	struct indoc_instructions *settings;
+	struct settings_block *settings;
 } cl_state;
 
-void Configuration::read_command_line(int argc, char **argv, indoc_instructions *settings) {
-	pathname *IM = Pathnames::from_text(I"indoc");
-	IM = Pathnames::subfolder(IM, I"Materials");
-	filename *basics = Filenames::in_folder(IM, I"basic-instructions.txt");
-	Configuration::add_instructions_file(basics);
+void Configuration::read_command_line(int argc, char **argv, settings_block *settings) {
 	cl_state state;
 	state.target_chosen = Str::new();
 	state.settings = settings;
 	@<Read the command line@>;
+	Configuration::add_instructions_file(
+		Filenames::in_folder(path_to_indoc_materials, I"basic-instructions.txt"));
 	Configuration::add_instructions_file(
 		Filenames::in_folder(settings->book_folder, I"indoc-instructions.txt"));
 	Configuration::read_instructions(state.target_chosen, settings);
@@ -73,7 +71,7 @@ void Configuration::read_command_line(int argc, char **argv, indoc_instructions
 
 @ =
 void Configuration::switch(int id, int val, text_stream *arg, void *v_cl_state) {
-	indoc_instructions *settings = ((cl_state *) v_cl_state)->settings;
+	settings_block *settings = ((cl_state *) v_cl_state)->settings;
 	switch (id) {
 		case VERBOSE_CLSW: settings->verbose_mode = val; break;
 		case TEST_INDEX_CLSW: settings->test_index_mode = val; break;
diff --git a/indoc/Chapter 1/Instructions.w b/indoc/Chapter 1/Instructions.w
index f9c81a8c8..b6af94b14 100644
--- a/indoc/Chapter 1/Instructions.w	
+++ b/indoc/Chapter 1/Instructions.w	
@@ -3,8 +3,10 @@
 Instructions of indoc to different output types.
 
 @h Definitions.
-
-@ A few fundamental options set at the command line.
+The command-line and Instructions-file-set values provide a large slate of
+what used to be global variables in the Perl version of Indoc. Today they
+are herded together into an instance of the |settings_block| structure,
+and in particular into a global instance of this called |indoc_settings|.
 
 @d LETTER_ALPHABETIZATION 1
 @d WORD_ALPHABETIZATION 2
@@ -29,7 +31,7 @@ Instructions of indoc to different output types.
 @d WRAPPER_zip 3
 
 =
-typedef struct indoc_instructions {
+typedef struct settings_block {
 	int verbose_mode;
 	int test_index_mode;
 
@@ -85,13 +87,13 @@ typedef struct indoc_instructions {
 	struct navigation_design *navigation;
 
 	MEMORY_MANAGEMENT
-} indoc_instructions;
+} settings_block;
 
 @
 
 =
-indoc_instructions *Instructions::clean_slate(void) {
-	indoc_instructions *settings = CREATE(indoc_instructions);
+settings_block *Instructions::clean_slate(void) {
+	settings_block *settings = CREATE(settings_block);
 	settings->verbose_mode = FALSE;
 	settings->test_index_mode = FALSE;
 
@@ -145,18 +147,11 @@ indoc_instructions *Instructions::clean_slate(void) {
 	settings->wrapper = WRAPPER_none;
 	settings->ebook = NULL;
 
-	settings->navigation = Gadgets::default();
+	settings->navigation = Nav::default();
 
 	return settings;
 }
 
-@ =
-typedef struct dc_metadatum {
-	struct text_stream *dc_key;
-	struct text_stream *dc_val;
-	MEMORY_MANAGEMENT
-} dc_metadatum;
-
 @h Instructions file.
 Note that |indoc| reports errors in the instructions file, but doesn't halt on
 them until all have been found. (The user may as well get all of the bad news,
@@ -164,13 +159,12 @@ not just the beginning of it.)
 
 =
 void Instructions::read_instructions(text_stream *target_sought, linked_list *L,
-	indoc_instructions *settings) {
+	settings_block *settings) {
 	int found_flag = FALSE; /* was a target of this name actually found? */
 
 	settings->change_logs_folder = Pathnames::subfolder(settings->book_folder, I"Change Logs");
 	settings->examples_directory = Pathnames::subfolder(settings->book_folder, I"Examples");
-	pathname *Materials = Pathnames::subfolder(Pathnames::from_text(I"indoc"), I"Materials");
-	settings->css_source_file = Filenames::in_folder(Materials, I"base.css");
+	settings->css_source_file = Filenames::in_folder(path_to_indoc_materials, I"base.css");
 	settings->definitions_index_leafname = Str::duplicate(I"general_index.html");
 
 	filename *F;
@@ -181,7 +175,7 @@ void Instructions::read_instructions(text_stream *target_sought, linked_list *L,
 	@<Reconcile any conflicting instructions@>;
 	@<Declare the format and wrapper as symbols@>;
 
-	HTMLUtilities::add_image_source(Pathnames::subfolder(Materials, I"images"));
+	HTMLUtilities::add_image_source(Pathnames::subfolder(path_to_indoc_materials, I"images"));
 
 	if (found_flag == FALSE)
 		Errors::fatal_with_text("unknown target %S", target_sought);
@@ -201,13 +195,13 @@ applies 20 for all targets except |hypercard|, where it applies 40.
 =
 typedef struct ins_helper_state {
 	int found_aim;
-	struct indoc_instructions *settings;
+	struct settings_block *settings;
 	struct text_stream *desired_target;
 	struct text_stream *scanning_target;
 } ins_helper_state;
 
 int Instructions::read_instructions_from(filename *F, text_stream *desired,
-	indoc_instructions *settings) {
+	settings_block *settings) {
 	ins_helper_state ihs;
 	ihs.scanning_target = Str::new();
 	ihs.desired_target = desired;
@@ -222,7 +216,7 @@ int Instructions::read_instructions_from(filename *F, text_stream *desired,
 void Instructions::read_instructions_helper(text_stream *cl, text_file_position *tfp,
 	void *v_ihs) {
 	ins_helper_state *ihs = (ins_helper_state *) v_ihs;
-	indoc_instructions *settings = ihs->settings;
+	settings_block *settings = ihs->settings;
 	match_results mr = Regexp::create_mr();
 
 	if (Regexp::match(&mr, cl, L" *#%c*")) { Regexp::dispose_of(&mr); return; }
@@ -269,9 +263,7 @@ void Instructions::read_instructions_helper(text_stream *cl, text_file_position
 		settings->book_contains_examples = TRUE;
 	} else if (Regexp::match(&mr, cl, L" *dc:(%C+): *(%c*?) *")) {
 		@<Disallow this in a specific target@>;
-		dc_metadatum *dcm = CREATE(dc_metadatum);
-		dcm->dc_key = Str::duplicate(mr.exp[0]);
-		dcm->dc_val = Str::duplicate(mr.exp[1]);
+		Instructions::create_ebook_metadata(Str::duplicate(mr.exp[0]), Str::duplicate(mr.exp[1]));
 	} else if (Regexp::match(&mr, cl, L" *css: *(%c*?) *")) {
 		@<Act on a CSS tweak@>;
 	} else if (Regexp::match(&mr, cl, L" *index: *(%c*?) *")) {
@@ -397,15 +389,20 @@ taste). In a multiple-line value, each line is terminated with a newline.
 
 @<Set an instructions option@> =
 	if (Str::eq_wide_string(key, L"alphabetization")) {
-		if (Str::eq_wide_string(val, L"word-by-word")) { settings->index_alphabetisation_algorithm = WORD_ALPHABETIZATION; }
-		else if (Str::eq_wide_string(val, L"letter-by-letter")) { settings->index_alphabetisation_algorithm = LETTER_ALPHABETIZATION; }
+		if (Str::eq_wide_string(val, L"word-by-word"))
+			settings->index_alphabetisation_algorithm = WORD_ALPHABETIZATION;
+		else if (Str::eq_wide_string(val, L"letter-by-letter"))
+			settings->index_alphabetisation_algorithm = LETTER_ALPHABETIZATION;
 		else Errors::in_text_file("no such alphabetization", tfp);
 	}
-	else if (Str::eq_wide_string(key, L"assume_Public_Library")) {
-		settings->assume_Public_Library = Instructions::set_yn(key, val, tfp); }
-	else if (Str::eq_wide_string(key, L"change_logs_directory")) { settings->change_logs_folder = Instructions::set_path(val, settings); }
-	else if (Str::eq_wide_string(key, L"contents_leafname")) { settings->contents_leafname = Str::duplicate(val); }
-	else if (Str::eq_wide_string(key, L"contents_expandable")) { settings->contents_expandable = Instructions::set_yn(key, val, tfp); }
+	else if (Str::eq_wide_string(key, L"assume_Public_Library"))
+		settings->assume_Public_Library = Instructions::set_yn(key, val, tfp);
+	else if (Str::eq_wide_string(key, L"change_logs_directory"))
+		settings->change_logs_folder = Instructions::set_path(val, settings);
+	else if (Str::eq_wide_string(key, L"contents_leafname"))
+		settings->contents_leafname = Str::duplicate(val);
+	else if (Str::eq_wide_string(key, L"contents_expandable"))
+		settings->contents_expandable = Instructions::set_yn(key, val, tfp);
 	else if (Str::eq_wide_string(key, L"css_source_file")) { settings->css_source_file = Instructions::set_file(val, settings); }
 	else if (Str::eq_wide_string(key, L"definitions_filename")) { settings->definitions_filename = Instructions::set_file(val, settings); }
 	else if (Str::eq_wide_string(key, L"definitions_index_filename")) {
@@ -452,7 +449,7 @@ taste). In a multiple-line value, each line is terminated with a newline.
 		settings->link_to_extensions_index = Str::duplicate(val); }
 	else if (Str::eq_wide_string(key, L"manifest_leafname")) { settings->manifest_leafname = Str::duplicate(val); }
 	else if (Str::eq_wide_string(key, L"navigation")) {
-		settings->navigation = Gadgets::parse(val);
+		settings->navigation = Nav::parse(val);
 		if (settings->navigation == NULL) Errors::in_text_file("no such navigation mode", tfp);
 	}
 	else if (Str::eq_wide_string(key, L"retina_images")) {
@@ -487,7 +484,7 @@ taste). In a multiple-line value, each line is terminated with a newline.
 		}
 		settings->contents_expandable = FALSE;
 		settings->images_copy = 1;
-		settings->navigation = Gadgets::for_ebook(settings->navigation);
+		settings->navigation = Nav::for_ebook(settings->navigation);
 		settings->format = HTML_FORMAT;
 		settings->XHTML = TRUE;
 		settings->ebook = Epub::new(I"untitled ebook", "");
@@ -511,7 +508,7 @@ taste). In a multiple-line value, each line is terminated with a newline.
 	}
 
 	if (settings->format == PLAIN_FORMAT)
-		settings->navigation = Gadgets::for_plain_text(settings->navigation);
+		settings->navigation = Nav::for_plain_text(settings->navigation);
 
 @<Declare the format and wrapper as symbols@> =
 	if (settings->wrapper == WRAPPER_epub) Symbols::declare_symbol(I"EPUB");
@@ -526,7 +523,7 @@ Note the Unix-style conveniences for pathnames: an initial |~| means the
 home folder, |~~| means the book folder.
 
 =
-pathname *Instructions::set_path(text_stream *val, indoc_instructions *settings) {
+pathname *Instructions::set_path(text_stream *val, settings_block *settings) {
 	if (Str::get_at(val, 0) == '~') {
 		if (Str::get_at(val, 1) == '~') {
 			if ((Str::get_at(val, 2) == '/') || (Str::get_at(val, 2) == FOLDER_SEPARATOR)) {
@@ -549,7 +546,7 @@ pathname *Instructions::set_path(text_stream *val, indoc_instructions *settings)
 }
 
 @ =
-filename *Instructions::set_file(text_stream *val, indoc_instructions *settings) {
+filename *Instructions::set_file(text_stream *val, settings_block *settings) {
 	if (Str::get_at(val, 0) == '~') {
 		if (Str::get_at(val, 1) == '~') {
 			if ((Str::get_at(val, 2) == '/') || (Str::get_at(val, 2) == FOLDER_SEPARATOR)) {
@@ -605,6 +602,18 @@ int Instructions::set_yn(text_stream *key, text_stream *val, text_file_position
 @ For ebooks only.
 
 =
+typedef struct dc_metadatum {
+	struct text_stream *dc_key;
+	struct text_stream *dc_val;
+	MEMORY_MANAGEMENT
+} dc_metadatum;
+
+void Instructions::create_ebook_metadata(text_stream *key, text_stream *value) {
+	dc_metadatum *dcm = CREATE(dc_metadatum);
+	dcm->dc_key = Str::duplicate(key);
+	dcm->dc_val = Str::duplicate(value);
+}
+
 void Instructions::apply_ebook_metadata(ebook *E) {
 	dc_metadatum *dcm;
 	LOOP_OVER(dcm, dc_metadatum) {
diff --git a/indoc/Chapter 1/Main.w b/indoc/Chapter 1/Main.w
index 6840b7e5d..b1890db6b 100644
--- a/indoc/Chapter 1/Main.w	
+++ b/indoc/Chapter 1/Main.w	
@@ -2,39 +2,15 @@
 
 The top level of the program.
 
-@h Definitions.
-
-@ Indoc is one of the earliest Inform tools, and it spent much of its
-life as a hacky Perl script: like all too many quick-fix Perl scripts, it
-was still in use ten years later. In 2012, I spent some time tidying it up
-to generate better HTML, and moved it over to literate code. This took an
-exasperatingly long time, not least because the original had produced
-typically sloppy turn-of-the-century HTML, with tables for layout and
-no CSS, and with many now-deprecated tags and elements. The 2012 edition,
-by contrast, needed to produce validatable XHTML 1.1 Strict in order to
-make EPUBs which read roughly correctly in today's ebook-readers, and
-when they call this Strict they're not kidding. It took something like
-four weeks of spare evenings.
-
-Just as I was finishing up, the programmer and commentator John Siracusa
-described an almost identical web-content-and-ebook generation task on his
-podcast (Hypercritical 85): "My solution for this is... I was trying to
-think of a good analogy for what happens when you're a programmer and you
-have this sort of task in front of you. Is it, the cobbler's children have
-no shoes? ... You would expect someone who is a programmer to make some
-awesome system which would generate these three things. But when you're a
-programmer, you have the ability to do whatever you want really, really
-quickly in the crappiest possible way... And that's what I did. I wrote a
-series of incredibly disgusting Perl scripts."
-
-This made me feel better. Nevertheless, in 2016, I rewrote in C.
-
 @h Nutshell.
 We turn the source matter, "rawtext", into a batch of output files using the
 chosen format, a process we'll call "rendering". We do this in two passes.
 
 =
-indoc_instructions *indoc_settings = NULL;
+pathname *path_to_indoc = NULL; /* where we are installed */
+pathname *path_to_indoc_materials = NULL; /* the materials pathname */
+
+settings_block *indoc_settings = NULL;
 int no_volumes = 0;
 int no_examples = 0;
 
@@ -61,10 +37,13 @@ int main(int argc, char **argv) {
 
 @<Start up indoc@> =
 	PRINT("indoc [[Version Number]] (Inform Tools Suite)\n");
-	Gadgets::start();
+	Nav::start();
 	Symbols::start_up_symbols();
 	indoc_settings = Instructions::clean_slate();
 	Configuration::read_command_line(argc, argv, indoc_settings);
+	path_to_indoc = Pathnames::installation_path("INDOC_PATH", I"indoc");
+	if (indoc_settings->verbose_mode) PRINT("Installation path is %p\n", path_to_indoc);
+	path_to_indoc_materials = Pathnames::subfolder(path_to_indoc, I"Materials");
 	if (indoc_settings->wrapper == WRAPPER_epub) {
 		HTMLUtilities::image_URL(NULL,
 			Filenames::get_leafname(indoc_settings->book_cover_image));
@@ -97,7 +76,7 @@ section by section.
 	volume *V;
 	text_stream *TO = NULL;
 	LOOP_OVER(V, volume) TO = Rawtext::process_large_rawtext_file(TO, V);
-	Gadgets::render_navigation_contents_files();
+	Nav::render_navigation_contents_files();
 
 @ The following functions here are for use only when compiling documentation
 to go inside the Inform user interface application.
diff --git a/indoc/Chapter 2/CSS.w b/indoc/Chapter 2/CSS.w
index 2de5f5562..4f7646d92 100644
--- a/indoc/Chapter 2/CSS.w	
+++ b/indoc/Chapter 2/CSS.w	
@@ -213,7 +213,7 @@ all CSS files, but here we know that |base.css| is tidily written.
 @ =
 void CSS::expand_IMAGES(text_stream *text) {
 	match_results mr = Regexp::create_mr();
-	if (indoc_settings->suppress_fonts == 1) {
+	if (indoc_settings->suppress_fonts) {
 		while (Regexp::match(&mr, text, L"(%c*?)font-family:(%c*?);(%c*)")) {
 			text_stream *L = mr.exp[0], *M = NULL, *R = mr.exp[2];
 			if (Regexp::match(NULL, mr.exp[1], L"%c*monospace%c*")) M = I"___MONOSPACE___";
diff --git a/indoc/Chapter 2/Renderer.w b/indoc/Chapter 2/Renderer.w
index 6d8234a88..e624b55f9 100644
--- a/indoc/Chapter 2/Renderer.w	
+++ b/indoc/Chapter 2/Renderer.w	
@@ -85,11 +85,11 @@ routine which doesn't surround the text with navigational gadgets and headings.
 text_stream *Renderer::render_block(OUTPUT_STREAM, volume *V, section *S) {
 	OUT = Renderer::formatted_file_must_be(OUT, V, S);
 
- 	Gadgets::render_navigation_top(OUT, V, S);
+ 	Nav::render_navigation_top(OUT, V, S);
  	Renderer::render_text_of_block(OUT, V, S);
- 	Gadgets::render_navigation_middle(OUT, V, S);
+ 	Nav::render_navigation_middle(OUT, V, S);
  	@<Render the examples below the text of the block@>;
- 	Gadgets::render_navigation_bottom(OUT, V, S);
+ 	Nav::render_navigation_bottom(OUT, V, S);
  	return OUT;
 }
 
@@ -102,14 +102,14 @@ text_stream *Renderer::render_block(OUTPUT_STREAM, volume *V, section *S) {
  		if (E->example_displayed_at_section[V->allocation_id] == S) {
 			no_examples_rendered_here++;
 			if (no_examples_rendered_here == 1)
-				Gadgets::render_navigation_example_top(OUT, V, S);
+				Nav::render_navigation_example_top(OUT, V, S);
 			@<Render the example here@>;
 			if (indoc_settings->examples_mode == EXMODE_open_internal) HTMLUtilities::ruled_line(OUT);
  		}
  	}
 
  	if (no_examples_rendered_here > 0)
- 		Gadgets::render_navigation_example_bottom(OUT, V, S);
+ 		Nav::render_navigation_example_bottom(OUT, V, S);
 
 @ Examples need to connect with particular sections of documentation, but
 they do so by title, not by block number, to protect them from renumbering
diff --git a/indoc/Chapter 3/Indexing Utilities.w b/indoc/Chapter 3/Indexing Utilities.w
index de85e3f2a..2df436806 100644
--- a/indoc/Chapter 3/Indexing Utilities.w	
+++ b/indoc/Chapter 3/Indexing Utilities.w	
@@ -35,7 +35,7 @@ text_stream *IndexUtilities::open_page(text_stream *title, text_stream *leafname
 	}
 	DISCARD_TEXT(head);
 
-	Gadgets::render_navigation_index_top(OUT, leafname, title);
+	Nav::render_navigation_index_top(OUT, leafname, title);
 	return OUT;
 }
 
diff --git a/indoc/Chapter 4/Codename Architect.w b/indoc/Chapter 4/Codename Architect.w
index 8ea368253..c980cfda7 100644
--- a/indoc/Chapter 4/Codename Architect.w	
+++ b/indoc/Chapter 4/Codename Architect.w	
@@ -8,7 +8,7 @@ heading includes these anyway.
 
 =
 navigation_design *Architect::create(void) {
-	navigation_design *ND = Gadgets::new(I"architect", FALSE, FALSE);
+	navigation_design *ND = Nav::new(I"architect", FALSE, FALSE);
 	ND->columnar = TRUE;
 	ND->contents_body_class = I"paper architectpapertint";
 	METHOD_ADD(ND, RENDER_SECTION_TITLE_MTID, Architect::architect_section_title);
diff --git a/indoc/Chapter 4/Codename Lacuna.w b/indoc/Chapter 4/Codename Lacuna.w
index 5587fe071..35023734a 100644
--- a/indoc/Chapter 4/Codename Lacuna.w	
+++ b/indoc/Chapter 4/Codename Lacuna.w	
@@ -7,7 +7,7 @@ no navigational features at all.
 
 =
 navigation_design *Lacuna::create(void) {
-	navigation_design *ND = Gadgets::new(I"lacuna", FALSE, TRUE);
+	navigation_design *ND = Nav::new(I"lacuna", FALSE, TRUE);
 	METHOD_ADD(ND, RENDER_CHAPTER_TITLE_MTID, Lacuna::lacuna_chapter_title);
 	METHOD_ADD(ND, RENDER_SECTION_TITLE_MTID, Lacuna::lacuna_section_title);
 	return ND;
diff --git a/indoc/Chapter 4/Codename Midnight.w b/indoc/Chapter 4/Codename Midnight.w
index c60225dd6..a99623631 100644
--- a/indoc/Chapter 4/Codename Midnight.w	
+++ b/indoc/Chapter 4/Codename Midnight.w	
@@ -6,7 +6,7 @@ The "midnight" style of navigational gadgets.
 
 =
 navigation_design *Midnight::create(void) {
-	navigation_design *ND = Gadgets::new(I"midnight", FALSE, FALSE);
+	navigation_design *ND = Nav::new(I"midnight", FALSE, FALSE);
 	ND->columnar = TRUE;
 	METHOD_ADD(ND, RENDER_SECTION_TITLE_MTID, Midnight::midnight_section_title);
 	METHOD_ADD(ND, RENDER_INDEX_TOP_MTID, Midnight::midnight_navigation_index_top);
@@ -197,7 +197,7 @@ void Midnight::write_contents_page(navigation_design *self, volume *V) {
 	TEMPORARY_TEXT(title);
 	WRITE_TO(title, "Contents");
 	@<Begin the HTML page for the contents@>;
-	Gadgets::navigation_contents_heading(OUT, V);
+	Nav::navigation_contents_heading(OUT, V);
 
 	for (int column = 0; column < no_volumes; column++)
 		if ((column == V->allocation_id) || (self->columnar))
diff --git a/indoc/Chapter 4/Codename Roadsign.w b/indoc/Chapter 4/Codename Roadsign.w
index 30bb9682c..b3b6e9a78 100644
--- a/indoc/Chapter 4/Codename Roadsign.w	
+++ b/indoc/Chapter 4/Codename Roadsign.w	
@@ -6,7 +6,7 @@ The "roadsign" style of navigational gadgets.
 
 =
 navigation_design *Roadsign::create(void) {
-	navigation_design *ND = Gadgets::new(I"roadsign", TRUE, FALSE);
+	navigation_design *ND = Nav::new(I"roadsign", TRUE, FALSE);
 	METHOD_ADD(ND, RENDER_VOLUME_TITLE_MTID, Roadsign::roadsign_volume_title);
 	METHOD_ADD(ND, RENDER_CHAPTER_TITLE_MTID, Roadsign::roadsign_chapter_title);
 	METHOD_ADD(ND, RENDER_SECTION_TITLE_MTID, Roadsign::roadsign_section_title);
diff --git a/indoc/Chapter 4/Codename Twilight.w b/indoc/Chapter 4/Codename Twilight.w
index 5e36854da..6a8c8e202 100644
--- a/indoc/Chapter 4/Codename Twilight.w	
+++ b/indoc/Chapter 4/Codename Twilight.w	
@@ -7,7 +7,7 @@ of "midnight".
 
 =
 navigation_design *Twilight::create(void) {
-	navigation_design *ND = Gadgets::new(I"twilight", FALSE, FALSE);
+	navigation_design *ND = Nav::new(I"twilight", FALSE, FALSE);
 	ND->simplified_examples = TRUE;
 	ND->simplified_letter_rows = TRUE;
 	METHOD_ADD(ND, RENDER_CHAPTER_TITLE_MTID, Twilight::twilight_chapter_title);
diff --git a/indoc/Chapter 4/Codename Unsigned.w b/indoc/Chapter 4/Codename Unsigned.w
index b5e318fa0..9bcfaafcb 100644
--- a/indoc/Chapter 4/Codename Unsigned.w	
+++ b/indoc/Chapter 4/Codename Unsigned.w	
@@ -6,7 +6,7 @@ The "unsigned" style of navigational gadgets.
 
 =
 navigation_design *Unsigned::create(void) {
-	navigation_design *ND = Gadgets::new(I"unsigned", TRUE, FALSE);
+	navigation_design *ND = Nav::new(I"unsigned", TRUE, FALSE);
 	METHOD_ADD(ND, RENDER_VOLUME_TITLE_MTID, Unsigned::unsigned_volume_title);
 	METHOD_ADD(ND, RENDER_CHAPTER_TITLE_MTID, Unsigned::unsigned_chapter_title);
 	METHOD_ADD(ND, RENDER_SECTION_TITLE_MTID, Unsigned::unsigned_section_title);
diff --git a/indoc/Chapter 4/Navigational Gadgets.w b/indoc/Chapter 4/Navigational Gadgets.w
deleted file mode 100644
index 29a275601..000000000
--- a/indoc/Chapter 4/Navigational Gadgets.w	
+++ /dev/null
@@ -1,221 +0,0 @@
-[Gadgets::] Navigational Gadgets.
-
-To render linking gadgets in HTML forms of documentation, so that
-the reader can navigate from section to section.
-
-@h
-
-=
-typedef struct navigation_design {
-	struct text_stream *codename;
-	int ebook_friendly;
-	int plain_friendly;
-	int columnar;
-	int simplified_examples;
-	int simplified_letter_rows;
-	struct text_stream *contents_body_class;
-	METHOD_CALLS
-	MEMORY_MANAGEMENT
-} navigation_design;
-
-navigation_design *Gadgets::new(text_stream *code, int e, int p) {
-	navigation_design *ND = CREATE(navigation_design);
-	ND->codename = Str::duplicate(code);
-	ND->ebook_friendly = e;
-	ND->plain_friendly = p;
-	ND->columnar = FALSE;
-	ND->simplified_examples = FALSE;
-	ND->simplified_letter_rows = FALSE;
-	ND->contents_body_class = I"paper midnightpapertint";
-	ND->methods = Methods::new_set();
-	return ND;
-}
-
-void Gadgets::start(void) {
-	Midnight::create(); /* needs to be created first */
-	Twilight::create();
-	Architect::create();
-	Roadsign::create(); /* needs to be created before unsigned */
-	Unsigned::create();
-	Lacuna::create();
-}
-
-navigation_design *Gadgets::default(void) {
-	return FIRST_OBJECT(navigation_design);
-}
-
-navigation_design *Gadgets::for_ebook(navigation_design *current) {
-	if (current->ebook_friendly) return current;
-	navigation_design *ND;
-	LOOP_OVER(ND, navigation_design)
-		if (ND->ebook_friendly)
-			return ND;
-	return NULL;
-}
-
-navigation_design *Gadgets::for_plain_text(navigation_design *current) {
-	if (current->plain_friendly) return current;
-	navigation_design *ND;
-	LOOP_OVER(ND, navigation_design)
-		if (ND->plain_friendly)
-			return ND;
-	return NULL;
-}
-
-navigation_design *Gadgets::parse(text_stream *val) {
-	navigation_design *ND;
-	LOOP_OVER(ND, navigation_design)
-		if (Str::eq(val, ND->codename))
-			return ND;
-	return NULL;
-}
-
-@h Top.
-At the front end of a section, before any of its text.
-
-@e RENDER_VOLUME_TITLE_MTID
-@e RENDER_CHAPTER_TITLE_MTID
-@e RENDER_SECTION_TITLE_MTID
-
-=
-VMETHOD_TYPE(RENDER_VOLUME_TITLE_MTID, navigation_design *ND, text_stream *OUT, volume *V)
-VMETHOD_TYPE(RENDER_CHAPTER_TITLE_MTID, navigation_design *ND, text_stream *OUT, volume *V, chapter *C)
-VMETHOD_TYPE(RENDER_SECTION_TITLE_MTID, navigation_design *ND, text_stream *OUT, volume *V, chapter *C, section *S)
-
-void Gadgets::render_navigation_top(OUTPUT_STREAM, volume *V, section *S) {
-	if (V->sections[0] == S) VMETHOD_CALL(indoc_settings->navigation, RENDER_VOLUME_TITLE_MTID, OUT, V);
-
-	chapter *C = S->begins_which_chapter;
-	if (C) VMETHOD_CALL(indoc_settings->navigation, RENDER_CHAPTER_TITLE_MTID, OUT, V, C);
-
-	if (indoc_settings->html_for_Inform_application)
-		@<Write HTML comments giving the Inform user interface search assistance@>;
-
-	VMETHOD_CALL(indoc_settings->navigation, RENDER_SECTION_TITLE_MTID, OUT, V, C, S);
-}
-
-@<Write HTML comments giving the Inform user interface search assistance@> =
-	WRITE("\n");
-	TEMPORARY_TEXT(comment);
-	WRITE_TO(comment, "SEARCH TITLE \"%S\"", S->unlabelled_title);
-	HTML::comment(OUT, comment);
-	Str::clear(comment);
-	WRITE_TO(comment, "SEARCH SECTION \"%S\"", S->label);
-	HTML::comment(OUT, comment);
-	Str::clear(comment);
-	WRITE_TO(comment, "SEARCH SORT \"%S\"", S->sort_code);
-	HTML::comment(OUT, comment);
-	DISCARD_TEXT(comment);
-
-@h Index top.
-And this is a variant for index pages, such as the index of examples.
-
-@e RENDER_INDEX_TOP_MTID
-
-=
-VMETHOD_TYPE(RENDER_INDEX_TOP_MTID, navigation_design *ND, text_stream *OUT, text_stream *filename, text_stream *title)
-
-void Gadgets::render_navigation_index_top(OUTPUT_STREAM, text_stream *filename, text_stream *title) {
-	VMETHOD_CALL(indoc_settings->navigation, RENDER_INDEX_TOP_MTID, OUT, filename, title);
-}
-
-@h Middle.
-At the middle part, when the text is over, but before any example cues.
-
-@e RENDER_NAV_MIDDLE_MTID
-
-=
-VMETHOD_TYPE(RENDER_NAV_MIDDLE_MTID, navigation_design *ND, text_stream *OUT, volume *V, section *S)
-
-void Gadgets::render_navigation_middle(OUTPUT_STREAM, volume *V, section *S) {
-	VMETHOD_CALL(indoc_settings->navigation, RENDER_NAV_MIDDLE_MTID, OUT, V, S);
-}
-
-@h Example top.
-This is reached before the first example is rendered, provided at least
-one example will be:
-
-@e RENDER_EXAMPLE_TOP_MTID
-
-=
-VMETHOD_TYPE(RENDER_EXAMPLE_TOP_MTID, navigation_design *ND, text_stream *OUT, volume *V, section *S)
-
-void Gadgets::render_navigation_example_top(OUTPUT_STREAM, volume *V, section *S) {
-
-	if (indoc_settings->format == HTML_FORMAT) {
-		HTML::begin_div_with_class_S(OUT, I"bookexamples");
-		HTML_OPEN_WITH("p", "class=\"chapterheading\"");
-	}
-
-	if (indoc_settings->examples_granularity == CHAPTER_GRANULARITY) {
-		chapter *C = S->in_which_chapter;
-		WRITE("Examples from %S", C->chapter_full_title);
-	} else if (indoc_settings->examples_granularity == BOOK_GRANULARITY) {
-		WRITE("Examples");
-	}
-
-	if (indoc_settings->format == HTML_FORMAT) {
-		HTML_CLOSE("p");
-	} else { WRITE("\n\n"); }
-
-	VMETHOD_CALL(indoc_settings->navigation, RENDER_EXAMPLE_TOP_MTID, OUT, V, S);
-}
-
-@h Example bottom.
-Any closing ornament at the end of examples? This is reached after the
-last example is rendered, provided at least one example has been.
-
-@e RENDER_EXAMPLE_BOTTOM_MTID
-
-=
-VMETHOD_TYPE(RENDER_EXAMPLE_BOTTOM_MTID, navigation_design *ND, text_stream *OUT, volume *V, section *S)
-
-void Gadgets::render_navigation_example_bottom(OUTPUT_STREAM, volume *V, section *S) {
-	if (indoc_settings->format == PLAIN_FORMAT) {
-		WRITE("\n\n");
-	}
-
-	if (indoc_settings->format == HTML_FORMAT) {
-		if (indoc_settings->examples_mode != EXMODE_open_internal) { HTMLUtilities::ruled_line(OUT); }
-		HTML::end_div(OUT);
-	}
-
-	VMETHOD_CALL(indoc_settings->navigation, RENDER_EXAMPLE_BOTTOM_MTID, OUT, V, S);
-}
-
-@h Bottom.
-At the end of the section, after any example cues and perhaps also example
-bodied. (In a section with no examples, this immediately follows the middle.)
-
-@e RENDER_NAV_BOTTOM_MTID
-
-=
-VMETHOD_TYPE(RENDER_NAV_BOTTOM_MTID, navigation_design *ND, text_stream *OUT, volume *V, section *S)
-
-void Gadgets::render_navigation_bottom(OUTPUT_STREAM, volume *V, section *S) {
-	if (indoc_settings->format == HTML_FORMAT) {
-		HTML::comment(OUT, I"START IGNORE");
-	}
-	VMETHOD_CALL(indoc_settings->navigation, RENDER_NAV_BOTTOM_MTID, OUT, V, S);
-	if (indoc_settings->format == HTML_FORMAT) {
-		HTML::comment(OUT, I"END IGNORE");
-	}
-}
-
-@h Contents page.
-Midnight provides a contents page of its very own.
-
-@e RENDER_CONTENTS_MTID
-@e RENDER_CONTENTS_HEADING_MTID
-
-=
-VMETHOD_TYPE(RENDER_CONTENTS_MTID, navigation_design *ND)
-VMETHOD_TYPE(RENDER_CONTENTS_HEADING_MTID, navigation_design *ND, text_stream *OUT, volume *V)
-
-void Gadgets::render_navigation_contents_files(void) {
-	VMETHOD_CALLV(indoc_settings->navigation, RENDER_CONTENTS_MTID);
-}
-
-void Gadgets::navigation_contents_heading(OUTPUT_STREAM, volume *V) {
-	VMETHOD_CALL(indoc_settings->navigation, RENDER_CONTENTS_HEADING_MTID, OUT, V);
-}
diff --git a/indoc/Contents.w b/indoc/Contents.w
index 7083c12c6..a07b6c6a6 100644
--- a/indoc/Contents.w
+++ b/indoc/Contents.w
@@ -12,11 +12,16 @@ Build Date: 16 February 2019
 
 Import: foundation
 
+Preliminaries
+	Introduction to Indoc
+	Volumes and Instructions
+	Documentation Markup
+
 Chapter 1: Setting Up
 	Basics
-	Instructions
-	Configuration
 	Main
+	Configuration
+	Instructions
 	Context Symbols
 
 Chapter 2: Processing
@@ -34,7 +39,7 @@ Chapter 3: Indexing
 	Examples Index
 
 Chapter 4: Navigation Styles
-	Navigational Gadgets
+	Navigational Designs
 	Codename Lacuna
 	Codename Midnight
 	Codename Architect
diff --git a/indoc/Preliminaries/Documentation Markup.w b/indoc/Preliminaries/Documentation Markup.w
new file mode 100644
index 000000000..64b6711e8
--- /dev/null
+++ b/indoc/Preliminaries/Documentation Markup.w	
@@ -0,0 +1,5 @@
+Documentation Markup.
+
+How to mark up the plain-text source for Inform documentation.
+
+@h Stuff.
diff --git a/indoc/Preliminaries/Introduction to Indoc.w b/indoc/Preliminaries/Introduction to Indoc.w
new file mode 100644
index 000000000..f72375845
--- /dev/null
+++ b/indoc/Preliminaries/Introduction to Indoc.w	
@@ -0,0 +1,93 @@
+Introduction to Indoc.
+
+What Indoc is, and its limited but complicated uses.
+
+@ Intest is a command line tool for generating (mainly) HTML or EPUB format
+documentation. A million of those have been written, and Indoc has no
+ambition to replace them. It is needed because Inform 7's documentation
+source consists of many small text files with idiosyncratic markup, while
+its formatted HTML version needs to be indexed in elaborate ways.
+
+Indoc is a purely command-line tool, used in building Inform but not in
+running it: it's not present in the Inform UI apps.
+
+If you have compiled the standard distribution of the command-line tools
+for Inform then the Indoc executable will be at |indoc/Tangled/indoc/|.
+Usage is very simple:
+
+	|$ indoc/Tangled/indoc [OPTIONS] TARGET|
+
+By default, Indoc reads its source documentation from a direction called
+|Documentation| (with respect to the current working directory); the
+option |-from X| changes this path to |X|, but in this manual we'll call
+it |Documentation|.
+
+In addition to documentation files, which will be described later, Indoc
+also reads instruction files. At minimum it will read
+
+	|Documentation/indoc-instructions.txt|
+
+but the option |-instructions X| causes it to read |X| as well. Instructions
+files mainly specify indexing notations, or CSS styles, or miscellaneous
+settings, but they group these under named "targets". For example:
+
+	|windows_app {|
+	|	...|
+	|}|
+
+declares a target called |windows_app|. (This is the form of HTML needed for
+use inside the Windows UI application for Inform.) The idea here is that
+there is probably no single form of HTML needed -- it will be needed in
+subtly different versions for different platforms: inside the app, as a
+stand-alone website, inside an Epub ebook. These different forms are
+called "targets". On any given run, Indoc generates a single target --
+the one named on the command line.
+
+The HTML produced is placed, by default, in the directory:
+
+	|Documentation/Output|
+
+This can be changed with the option |-to X|.
+
+@ When it runs, Indoc needs to know where it is installed in the file
+system. There is no completely foolproof, cross-platform way to know this
+(on some Unixes, a program cannot determine its own location), so Indoc
+decides by the following set of rules:
+
+(a) If the user, at the command line, specified |-at P|, for some path
+|P|, then we use that.
+(b) Otherwise, if the host operating system can indeed tell us where the
+executable is, we use that. This is currently implemented only on MacOS,
+Windows and Linux.
+(c) Otherwise, if the environment variable |$INDOC_PATH| exists and is
+non-empty, we use that.
+(d) And if all else fails, we assume that the location is |indoc|, with
+respect to the current working directory.
+
+If you're not sure what Indoc has decided and suspect it may be wrong,
+running Indoc with the |-verbose| switch will cause it to print its belief
+about its location as it starts up.
+
+@ As a program, Indoc began as a rat's nest of Perl in 2002, and you can still
+see where the rats used to live. Like all too many quick-fix Perl scripts, it
+was still in use ten years later. In 2012, I spent some time tidying it up to
+generate better HTML, and made it a web (that is, a literate program). The
+original had produced typically sloppy turn-of-the-century HTML, with tables
+for layout and no CSS, and with many now-deprecated tags and elements. The
+2012 edition, by contrast, needed to produce validatable XHTML 1.1 Strict in
+order to make EPUBs which read roughly correctly in today's ebook-readers, and
+when they call this Strict they're not kidding. It took something like four
+weeks of spare evenings.
+
+Just as I was finishing up, the programmer and commentator John Siracusa
+described a not dissimilar web-content-and-ebook generation task on his
+podcast (Hypercritical 85): "I was trying to think of a good analogy for what
+happens when you're a programmer and you have this sort of task in front of
+you. Is it, the cobbler's children have no shoes? ... You would expect someone
+who is a programmer to make some awesome system which would generate these
+three things. But when you're a programmer, you have the ability to do
+whatever you want really, really quickly in the crappiest possible way... And
+that's what I did. I wrote a series of incredibly disgusting Perl scripts."
+
+This made me feel better. Nevertheless, in 2016, indoc was rewritten in C,
+and it received a further revision in 2019.
diff --git a/indoc/Preliminaries/Volumes and Instructions.w b/indoc/Preliminaries/Volumes and Instructions.w
new file mode 100644
index 000000000..878d1cdbb
--- /dev/null
+++ b/indoc/Preliminaries/Volumes and Instructions.w	
@@ -0,0 +1,321 @@
+Volumes and Instructions.
+
+Dual- versus single-volume mode, and how to write instructions files.
+
+@h Model.
+Conceptually, an Indoc project has either one or two volumes. The source for
+each volume is a single UTF-8 encoded plain text file. In the core Inform
+repository, there are two volumes, with the files being
+
+	|Documentation/Writing with Inform.txt|
+	|Documentation/The Recipe Book.txt|
+
+These are independent books, with individual titles. It would seem simpler
+just to make them two different Indoc projects, but in dual-volume mode,
+Indoc can generate joint contents pages, and provide crosswise HTML links
+between the two volumes.
+
+The project can also include a number of "Examples", each being a single
+text file such as:
+
+	|Documentation/Examples/Prague.txt|
+
+which is the source for an Inform example called "The Prague Job".
+(These same text files are also used by Intest to test that all of the code
+samples included in the Inform documentation actually work as claimed.)
+There can be any number of examples, including none; Inform currently has 468.
+
+Each volume is divided into a series of chapters, and each chapter into a
+series of sections. Examples are always placed at the ends of sections;
+note that in dual-volume mode, examples are (mostly) present in both volumes,
+giving them two different locations. Thus, "The Prague Job" appears in section
+"More general linkages" of chapter "Scenes" of volume "Writing with Inform",
+and also in section "Scripted Scenes" of chapter "Time and Plot" of volume
+"The Recipe Book".
+
+@h Project instructions.
+The main instructions file for an Indoc project is, as noted earlier, at:
+
+	|Documentation/indoc-instructions.txt|
+
+An instruction file is a UTF-8 encoded plain text file. Single instructions
+occupy single lines (i.e., line breaks are significant). A white-space line,
+or a line whose first non-white-space character is a |#|, are ignored.
+
+The file should begin by specifying one or two volumes, and then, if they
+will contain Examples in the above sense, by giving the special |examples|
+instruction. Inform opens thus:
+
+	|volume: Writing with Inform|
+	|volume: The Inform Recipe Book (RB) = The Recipe Book.txt|
+	|examples|
+
+But a simpler, single-volume project might have just:
+
+	|volume: Pandemonium 2.0 for Fun and Profit|
+
+Each volume has a title, and Indoc automatically generates an abbreviation
+for it: by default, it takes the capital letters from the title, so that it
+abbreviates "Writing with Inform" to WI. That same method would have made
+turned "The Inform Recipe Book" into TIRB, but because we didn't want that,
+we supplied our own abbreviation RB instead.
+
+The third, also optional, part of a |volume| instruction specifies the
+leafname of the documentation source file for it. By default, this will be
+the title plus |.txt|: for example, |Writing with Inform.txt|. But we can
+use |= X| to specify that it should be |X| instead.
+
+Two other project instructions exist:
+
+If the project will contain images, then they will be looked for in a list
+of places. Top of the list is a directory internal to Indoc which includes
+some navigation icons such as |arrow-up.png|. The instruction |images: X|
+adds the directory |X| to this source list.
+
+Lastly, the cover image for the project can be specified with an instruction
+such as:
+
+	|cover: combined_cover.png|
+
+This specifies a leafname which must exist in one of the image sources
+mentioned above.
+
+@h Durham Core metadata.
+If the project needs to generate Epub books, then these will need to have
+some basic DC ("Durham Core") metadata supplied. For example:
+
+	|dc:title: Inform - A Design System for Interactive Fiction|
+	|dc:creator: Graham Nelson and Emily Short|
+	|dc:subject: Interactive Fiction|
+	|dc:identifier: wwi-rb-combined|
+
+The instruction |dc:KEY: VALUE| supplies a DC key-value pair.
+
+@h Targets.
+The instructions file typically begins as above, but then goes into a
+block of general settings or instructions (for which see below); and
+eventually gets around to describing one or more targets. A target
+looks like so:
+
+	|IDENTIFIER {|
+	|	...|
+	|}|
+
+where |IDENTIFIER| is its name. Targets, as noted in the introduction,
+are different forms of the documentation we might need to produce: Inform,
+for example, has targets called |plain|, |website|, |linux_app| and so on.
+What's important here is not that these are written to different locations
+on disc (though they are) but that they have finicky little differences
+in settings. The |...| stretch of lines can specify these. For example:
+
+	|ebook {|
+	|	granularity = 2|
+	|	examples_mode = open|
+	|	follow: epub-css-tweaks.txt|
+	|}|
+
+makes two specific settings and one instruction, all applying only for the
+target |ebook|.
+
+@h Symbols.
+The instruction |declare: SYMBOL| creates the symbol |SYMBOL|. These exist
+so that we can mark certain paragraphs of documentation as being present in
+only some of the targets.
+
+For example, we might want Linux installation instructions to appear only
+in the Linux version of a manual. To do that, we'll need the symbol:
+
+	|linux_app {|
+	|	...|
+	|	declare: Linux|
+	|	...|
+	|}|
+
+In the documentation, we could then mark up a paragraph like so:
+
+	|{Linux:}To install, first...|
+
+The symbol |indoc| is always declared, but by default no other symbols are.
+Lastly, |undeclare: SYMBOL| removes a symbol.
+
+@h Other instructions.
+|follow: I| tells Indoc to follow the instructions file |I|. This works
+rather like |#include| in C, or similar languages. If the |follow:| is
+included inside a target block, then it affects only that target. On
+other targets, the file |I| won't even be opened, and need never exist.
+
+|css:| specifies additional CSS (Cascading Style Sheet) styling. This
+will be needed only if, for example, unusual indexing features are used,
+in which different categories of index entry need different visual styling.
+For example,
+
+	|css: span.indextitle ++ {|
+	|	font-style: italic;|
+	|}|
+
+Here the material between the braces is pure CSS, not Indoc syntax. The
+notation |++| here tells Indoc that an entirely new CSS style is being
+created; |+| would supply new lines to an existing style.
+
+|index: NOTATION = CATEGORY OPTION| defines a new indexing markup notation;
+for example,
+
+	|index: ^{@headword} = name (invert)|
+
+says that markup notations like |^{@Andrew Plotkin}| put a name into the index,
+which should be an index entry of category |name|, and should be inverted,
+in that it will be alphabetised under "Plotkin, Andrew". The text |headword|
+in the prototype is where the entry text should appear in the notation.
+
+@h Miscellaneous settings.
+There are a great many of these, but most are set to sensible defaults,
+and it is not compulsory to set any of them. Lines such as
+
+	|SETTING = VALUE|
+
+change the default settings if need be. Here is an A-Z list; they're really
+too miscellaneous to be grouped usefully by subject matter.
+
+|alphabetization| sets the index sorting algorithm. The default is
+|letter-by-letter|; the alternative is |word-by-word|. The difference is
+that letter-by-letter would ignore word divisions and sort in the order
+"peach", "peachpit", "peach tree"; whereas word-by-word would go for
+"peach", "peach tree", "peachpit". 
+
+|assume_Public_Library| can be |yes| or |no|. The default is |no|. This
+specifies whether special HTML links to the Public Library will be valid;
+outside of Inform UI apps, the answer is definitely no.
+
+|change_logs_directory| is the path to a directory holding Inform release
+change log files. By default, this will be |Documentation/Change Logs|.
+
+|contents_leafname| is the (unextended) leafname to give the HTML contents
+page. The default is |index|.
+
+|contents_expandable| can be |yes| or |no|. The default is |no|. This sets
+whether Javascript-powered "expand" buttons are to be used in the contents
+page, and has effect only on the Midnight navigation design.
+
+|css_source_file| is the filename of the CSS style sheet to use. The default
+is the |base.css| file included in the Indoc distribution.
+
+|definitions_filename| is the filename to use if you would like Indoc to
+output a special file of Inform phrase definitions, for use by Inform itself
+when it generates indexes. The default for this is |definitions.html|. This
+has nothing to do with the |definitions_index_filename|.
+
+|definitions_index_filename| is the leafname to use for the General Index
+in the documentation. The default is |general_index.html|. This
+has nothing to do with the |definitions_filename|.
+
+|destination| is the directory into which output is generated. The default
+is |Documentation/Output|. Note that specifying |-to X| at the command line
+overrides this setting: if |-to| is used, |destination| is ignored.
+
+|examples_directory| is the directory holding the Example files. The default
+is |Documentation/Examples|.
+
+|examples_alphabetical_leafname| is the leafname to use for the alphabetical
+index of examples in the documentation. The default is |examples_alphabetical.html|.
+
+|examples_granularity| is 1, 2, or 3. It can never be less than |granularity|,
+and by default is equal to it. It specifies where examples should appear:
+at the end of the relevant volume (1), chapter (2), or section (3).
+
+|examples_mode| is |open| or |openable|, and is by default |open|. Open means
+that an example has its full contents visible by default; openable means that
+the contents are hidden behind a Javascript-powered button which causes them
+to be revealed.
+
+|examples_numerical_leafname| is the leafname to use for the numerical
+index of examples in the documentation. The default is |examples_numerical.html|.
+
+|examples_thematic_leafname| is the leafname to use for the thematic
+index of examples in the documentation. The default is |examples_thematic.html|.
+
+|format| is the most important of all the settings, and is |HTML| or |text|,
+but by default |HTML| unless the target name is |plain|, in which case |text|.
+
+|granularity| is 1, 2, or 3. The default is 3 unless the target is called
+|webpage| or |plain|, in which case it is 1. This specifies how much the
+documentation is broken down into pieces. 1 means "each volume in a single
+HTML file"; 2 means "each chapter", 3 means "each section". Low granularity
+means fewer but larger files, high granularity more but smaller files.
+
+|html_for_Inform_application| can be |yes| or |no|. The default is |no|. This
+specifies whether the HTML is for use inside the Inform UI application, and
+can therefore use links with the special HTTP transports only available there.
+
+|images_copy| can be |yes| or |no|. The default is |yes|. In this mode,
+any needed image files are copied into place into the |images_path|. (The
+alternative assumes they are already there, and should be used if |images_path|
+is some URL external to the HTML being generated.)
+
+|images_path| is where the generated HTML expects to find its image files.
+The default is |~~/Images/|, where |~~| means the destination directory:
+that is, the default is a subdirectory called |Images| of the destination.
+
+|inform_definitions_mode| can be |yes| or |no|. The default is |no|. This
+is cosmetic, and provides extra styling on lines of documentation giving the
+syntax for Inform phrases.
+
+|javascript| can be |yes| or |no|. The default is |yes|. This indicates
+whetber Indoc is allowed to compile Javascript, or has to stick to inactive
+HTML.
+
+|javascript_paste_method| can be |none|, |Andrew| or |David|. The default
+is |none|. The difference relates to how "paste Inform source" links are
+implemented inside the Inform application: |Andrew| mode is suitable for
+most platforms, but |David| is needed for Windows.
+
+|link_to_extensions_index| is meaningful only if |html_for_Inform_application|
+is set, and specifies the URL of the Extensions index inside the app.
+
+|manifest_leafname| is meaningful only if |html_for_Inform_application|
+is set, and is by default |manifest.txt|. This provides a cross-reference
+list of files generated by Indoc.
+
+|navigation| is the design used for navigation links in the HTML produced.
+There are currently six designs, called |architect|, |lacuna|, |midnight|,
+|roadsign|, |twilight|, and |unsigned|; the default is |roadsign|, though
+inside the Inform applications, the design chosen is usually |architect|.
+If the format is |text| not |HTML|, then the design is always |lacuna|.
+
+|retina_images| can be |yes| or |no|. The default is |no|. This indicates
+whether MacOS/iOS "retina" versions of the paste and create icons are
+available: |paste@2x.png| and |create@2x.png| respectively.
+
+|support_creation| can be |yes| or |no|. The default is |no|. This indicates
+whether the Examples have a "create" button which creates a new Inform
+project demonstrating them in action; this can only be done in the UI apps,
+so it should always be |no| unless |html_for_Inform_application| is |yes|.
+
+|suppress_fonts| can be |yes| or |no|. The default is |no|. If |yes|, this
+strips out lists of fonts to use in CSS, leaving only whether they are
+|monospace| or not.
+
+|toc_granularity| is 1, 2, or 3. It can never be less than |granularity|,
+and by default is 3. It shows the level of detail in the table of contents: 1
+means volumes, 2 means volumes and chapters, 3 goes down to sections.
+
+|top_and_tail| specifies a prototype HTML file to follow for the more
+important HTML pages generated by Indoc. The default is not to. If
+this exists, it can provide a surround for the HTML we generate -- for
+example, it can contain website-specific navigation, or a banner across
+the top. The prototype should somewhere include the text |[TEXT]|, and
+this will be replaced with whatever Indoc generates.
+
+|top_and_tail_sections| is the same as |top_and_tail|, but for individual
+section files.
+
+|treat_code_as_verbatim| can be |yes| or |no|. The default is |yes|. This
+affects the styling of marked-up code material in documentation. Without
+it, code markup is largely unavailable.
+
+|wrapper| can be |EPUB|, |zip| or |none|. The default is |none|. The wrapper
+is put around the whole mass of generated HTML; |EPUB| makes the result an
+Epub-format ebook.
+
+|XHTML| can be |yes| or |no|. The default is |no|. This forces the HTML we
+produce to conform to XHTML 1.1 Strict. If the |wrapper| is |EPUB|, then
+this is automatically set to |yes|.