First instalment of a manual for Halibut. This lot covers the

[sgt/halibut] / doc / input.but

\C{input} Halibut input format

This chapter describes the format in which you should write
documents to be processed by Halibut.

\H{input-basics} The basics

Halibut's input files mostly look like ordinary ASCII text files;
you can edit them with any text editor you like.

Writing paragraphs of ordinary text is very simple: you just write
ordinary text in the ordinary way. You can wrap a paragraph across
more than one line using line breaks in the text file, and Halibut
will ignore this when it rewraps the paragraph for each output
format. To separate paragraphs, use a blank line (i.e. two
consecutive line breaks). For example, a fragment of Halibut input
looking like this:

\c This is a line of text.
\c This is another line of text.
\c
\c This line is separated from the previous one by a blank line.

will produce two paragraphs looking like this:

\quote{
This is a line of text.
This is another line of text.

This line is separated from the previous one by a blank line.
}

The first two lines of the input have been merged together into a
single paragraph, and the line break in the input file was treated
identically to the spaces between the individual words.

Halibut is designed to have very few special characters. The only
printable characters in Halibut input which will not be treated
exactly literally in the output are the backslash (\c{\\}) and the
braces (\c{\{} and \c{\}}). If you do not use these characters,
\e{everything} else you might type in normal ASCII text is perfectly
safe. If you do need to use any of those three characters in your
document, you will have to precede each one with a backslash. Hence,
for example, you could write

\c This \\ is a backslash, and these are \{braces\}.

and Halibut would generate the text

\quote{
This \\ is a backslash, and these are \{braces\}.
}

\H{input-inline} Simple inline formatting commands

Halibut formatting commands all begin with a backslash, followed by
a word or character identifying the command. Some of them then use
braces to surround one or more pieces of text acted on by the
command. (In fact, the \c{\\\\}, \c{\\\{} and \c{\\\}} sequences you
met in \k{input-basics} are themselves formatting commands.)

This section describes some simple formatting commands you can use
in Halibut documents. The commands in this section are \e{inline}
commands, which means you can use them in the middle of a paragraph.
\K{input-para} describes some \e{paragraph} commands, which affect a
whole paragraph at a time.

\S{input-emph} \c{\\e}: Emphasising text

Possibly the most obvious piece of formatting you might want to use
in a document is \e{emphasis}. To emphasise text, you use the
\c{\\e} command, and follow it up with the text to be emphasised in
braces. For example, the first sentence in this paragraph was
generated using the Halibut input

\c Possibly the most obvious piece of formatting you might want to use
\c in a document is \e{emphasis}.

\S{input-code} \c{\\c} and \c{\\cw}: Displaying computer code inline

Halibut was primarily designed to produce software manuals. It can
be used for other types of document as well, but software manuals
are its speciality.

In software manuals, you often want to format text in a way that
indicates that it is something you might see displayed verbatim on
a computer screen. In printed manuals, this is typically done by
setting that text in a font which is obviously fixed-width. This
provides a visual cue that the text being displayed is code, and it
also ensures that punctuation marks are clearly separated and shown
individually (so that a user can copy the text accurately and
conveniently).

Halibut provides \e{two} commands for this, which are subtly
different. The names of those commands are \c{\\c} (\q{code}) and
\c{\\cw} (\q{weak code}). You use them just like \c{\\e}, by
following them with some text in braces. For example, this...

\c This sentence contains some \c{code} and some \cw{weak code}.

... produces this:

\quote{
This sentence contains some \c{code} and some \cw{weak code}.
}

The distinction between code and weak code is mainly important when
producing plain text output. Plain text output is typically viewed
in a fixed-width font, so there is no need (and no way) to change
font in order to make the order of punctuation marks clear. However,
marking text as code is also \e{sometimes} done to provide a visual
distinction between it and the text around it, so that the reader
knows where the literal computer text starts and stops; and in plain
text, this cannot be done by changing font, so there needs to be an
alternative way.

So in the plain text back end, things marked as code (\c{\\c}) will
be surrounded by quote marks, so that it's obvious where they start
and finish. Things marked as weak code (\c{\\cw}) will not look any
different from normal text.

I recommend using weak code for any application where it is
\e{obvious} that the text is literal computer input or output. For
example, if the text is capitalised, that's usually good enough. If
I talk about the Pentium's \cw{EAX} and \cw{EDX} registers, for
example, you don't need quotes to notice that those are special; so
I would write that in Halibut as \q{\c{the Pentium's \\cw\{EAX\} and
\\cw\{EDX\} registers}}. But if I'm talking about the Unix command
\c{man}, which is an ordinary English word in its own right, a reader
might be slightly confused if it appeared in the middle of a
sentence undecorated; so I would write that as \q{\c{the Unix command
\\c\{man\}}}.

In summary:

\b \c{\\c} means \q{this text \e{must} be visually distinct from the
text around it}. Halibut's various output formats will do this by
changing the font if possible, or by using quotes if not.

\b \c{\\cw} means \q{it would be nice to display this text in a
fixed-width font if possible, but it's not essential}.

There is a separate mechanism for displaying computer code in an
entire paragraph; see \k{input-codepara} for that one.

\S{input-quotes} \c{\\q}: Quotation marks

Halibut's various output formats don't all use the same conventions
for displaying text in ordinary quotation marks (\q{like these}).
Some output formats have access to proper matched quote characters,
whereas others are restricted to using plain ASCII. Therefore, it is
not ideal to use the ordinary ASCII double quote character in your
document (although you can if you like).

Halibut provides the formatting command \c{\\q} to indicate quoted
text. If you write

\c Here is some \q{text in quotes}.

then Halibut will print

\quote{
Here is some \q{text in quotes}.
}

and in every output format Halibut generates, it will choose the
best quote characters available to it in that format.

You can still use ordinary ASCII double quotes if you prefer; or you
could even use the \c{\\u} command (see \k{input-unicode}) to
generate Unicode matched double quotes and fall back to the normal
ASCII one if they aren't available. But I recommend using the
built-in \c{\\q} command in most cases, because it's simple and does
the best it can everywhere.

(Note that if you're using the \c{\\c} or \c{\\cw} commands to
display literal computer code, you probably \e{will} want to use
literal ASCII quote characters, because it is likely to matter
precisely which quote character you use.)

\S{input-nonbreaking} \c{\\-} and \c{\\_}: Non-breaking hyphens and
spaces

If you use an ordinary hyphen in the middle of a word (such as
\q{built-in}), Halibut's output formats will feel free to break a
line after that hyphen when wrapping paragraphs. This is fine for a
word like \q{built-in}, but if you were displaying some literal
computer code such as the Emacs command
\c{M\-x\_psychoanalyze\-pinhead}, you might prefer to see the whole
hyphenated word treated as an unbreakable block. In some cases, you
might even want to prevent the \e{space} in that command from
becoming a line break.

For these purposes, Halibut provides the commands \c{\\-} and
\c{\\_}, which generate a non-breaking hyphen and a non-breaking
space respectively. So the above Emacs command might be written as

\c the Emacs command \c{M\-x\_psychoanalyze\-pinhead}

Unfortunately, some of Halibut's output formats do not support
non-breaking hyphens, and others don't support \e{breaking} hyphens!
So Halibut cannot promise to honour these commands in all situations.
All it can do is make a best effort.

\S{input-date} \c{\\date}: Automatic date generation

Sometimes you might want your document to give an up-to-date
indication of the date on which it was run through Halibut.

Halibut supplies the \c{\\date} command to do this. In its simplest
form, you simply say

\c This document was generated on \date.

and Halibut generates something like

\quote{
This document was generated on \date.
}

If you would prefer the date to be generated in a specific format,
you can follow the \c{\\date} command with a format specification in
braces. The format specification will be run through the standard C
function \c{strftime}, so any format acceptable to that function is
acceptable here as well. I won't document the format here, because
the details vary from computer to computer (although there is a
standard core which should be supported everywhere). You should look
at your local system's manual for \c{strftime} for details.

Here's an example which generates the date in the international
standard ISO 8601 format:

\c This document was generated on \date{%Y-%m-%d %H:%M:%S}.

And here's some sample output from that command:

\quote{
This document was generated on \date{%Y-%m-%d %H:%M:%S}.
}

\S{input-weblink} \c{\\W}: WWW hyperlinks

Since one of Halibut's output formats is HTML, it's obviously useful
to be able to provide links to arbitrary web sites in a Halibut
document.

This is done using the \c{\\W} command. \c{\\W} expects to be
followed by \e{two} sets of braces. In the first set of braces you
put a URL; in the second set you put the text which should be a
hyperlink. For example, you might write

\c Try searching on \W{http://www.google.com/}{Google}.

and Halibut would generate

\quote{
Try searching on \W{http://www.google.com/}{Google}.
}

Note that hyperlinks, like the non-breaking commands discussed in
\k{input-nonbreaking}, are \e{discretionary}: if an output format
does not support them then they will just be left out completely. So
unless you're \e{only} intending to use the HTML output format, you
should avoid storing vital content in the URL part of a \c{\\W}
command. The Google example above is reasonable (because most users
are likely to be able to find Google for themselves even without a
convenient hyperlink leading straight there), but if you really need
to direct users to a specific web site, you will need to give the
URL in actual displayed text (probably displayed as code as well).
However, there's nothing to stop you making it a hyperlink \e{as
well} for the convenience of HTML readers.

The \c{\\W} command supports a piece of extra syntax to make this
convenient for you. You can specify \c{\\c} or \c{\\cw} \e{between}
the first and second pairs of braces. For example, you might write

\c Google is located at \W{http://www.google.com/}\cw{www.google.com}.

and Halibut would produce

\quote{
Google is located at \W{http://www.google.com/}\cw{www.google.com}.
}

\S{input-unicode} \c{\\u}: Specifying arbitrary Unicode characters

When Halibut is finished, it should have full Unicode support. You
should be able to specify any (reasonably well known) character set
for your input document, and Halibut should convert it all to
Unicode as it reads it in. Similarly, you should be able to specify
the character set you want for each output format and have all the
conversion done automatically.

Currently, none of this is actually supported. Input text files are
assumed to be in ISO 8859-1, and each output format has its own
non-configurable character set (although the HTML output can use the
\c{Ӓ} mechanism to output any Unicode character it likes).

If you need to specify a Unicode character in your input document
which is not supported by the input character set, you can use the
\c{\\u} command to do this. \c{\\u} expects to be followed by a
sequence of hex digits; so that \c{\\u0041}, for example, denotes
the Unicode character \cw{0x0041}, which is the capital letter A.

If a Unicode character specified in this way is not supported in a
particular \e{output} format, you probably don't just want it to be
omitted. So you can put a pair of braces after the \c{\\u} command
containing fallback text. For example, to specify an amount of money
in euros, you might write this:

\c This is likely to cost \u20AC{EUR }2500 at least.

Halibut will render that as a Euro sign \e{if available}, and
the text \q{EUR } if not. In the output format you're currently
reading in, the above input generates this:

\quote{
This is likely to cost \u20AC{EUR }2500 at least.
}

If you read it in other formats, you may see different results.

\S{input-xref} \c{\\k} and \c{\\K}: Cross-references to other sections

\K{intro-features} mentions that Halibut numbers the sections of
your document automatically, and can generate cross-references to
them on request. \c{\\k} and \c{\\K} are the commands used to
generate those cross-references.

To use one of these commands, you simply follow it with a pair of
braces containing the keyword for the section in question. For
example, you might write something like

\c \K{input-xref} expands on \k{intro-features}.

and Halibut would generate something like

\quote{
\K{input-xref} expands on \k{intro-features}.
}

The keywords \c{input-xref} and \c{intro-features} are section
keywords used in this manual itself. In your own document, you would
have supplied a keyword for each one of your own sections, and you
would provide your own keywords for the \c{\\k} command to work on.

The difference between \c{\\k} and \c{\\K} is simply that \c{\\K}
starts the cross-reference text with a capital letter; so you would
use \c{\\K} at the beginning of a sentence, and \c{\\k} everywhere
else.

In output formats which permit it, cross-references act as
hyperlinks, so that clicking the mouse on a cross-reference takes
you straight to the referenced section.

The \c{\\k} commands are also used for referring to entries in a
bibliography (see \k{input-biblio} for more about bibliographies),
and can also be used for referring to an element of a numbered list
by its number (see \k{input-list-number} for more about numbered
lists).

See \k{input-sections} for more about chapters and sections.

\S{input-inline-comment} \c{\\#}: Inline comments

If you want to include comments in your Halibut input, to be seen
when reading it directly but not copied into the output text, then
you can use \c{\\#} to do this. If you follow \c{\\#} with text in
braces, that text will be ignored by Halibut.

For example, you might write

\c The typical behaviour of an antelope \#{do I mean gazelle?} is...

and Halibut will simply leave out the aside about gazelles, and will
generate nothing but

\quote{
The typical behaviour of an antelope \#{do I mean gazelle?} is...
}

The \c{\\#} command can also be used to produce a whole-paragraph
comment; see \k{input-commentpara} for details of that.

\H{input-para} Paragraph-level commands

This section describes Halibut commands which affect an entire
paragraph, or sometimes even \e{more} than one paragraph, at a time.

\S{input-codepara} \c{\\c}: Displaying whole paragraphs of computer
code

\K{input-code} describes a mechanism for displaying computer code in
the middle of a paragraph, a few words at a time.

However, this is often not enough. Often, in a computer manual, you
really want to show several lines of code in a display paragraph.

This is also done using the \c{\\c} command, in a slightly different
way. Instead of using it in the middle of a paragraph followed by
braces, you can use it at the start of each line of a paragraph. For
example, you could write

\c \c #include <stdio.h>
\c \c
\c \c int main(int argc, char **argv) {
\c \c     printf("hello, world\n");
\c \c     return 0;
\c \c }

and Halibut would generate

\quote{

\c #include <stdio.h>
\c
\c int main(int argc, char **argv) {
\c     printf("hello, world\n");
\c     return 0;
\c }

}

Note that the above paragraph makes use of a backslash and a pair of
braces, and does \e{not} need to escape them in the way described in
\k{input-basics}. This is because code paragraphs formatted in this
way are a special case; the intention is that you can just copy and
paste a lump of code out of another program, put \q{\cw{\\c }} at the
start of every line, and simply \e{not have to worry} about the
details - you don't have to go through the whole block looking for
characters to escape.

Since a backslash inside a code paragraph generates a literal
backslash, this means you cannot use any other Halibut formatting
commands inside a code paragraph. In particular, if you want to
emphasise a particular word in the paragraph, you can't do that
using \c{\\e} (\k{input-emph}) in the normal way.

Therefore, Halibut provides an alternative means of emphasis in code
paragraphs. Each line beginning with \c{\\c} can optionally be
followed by a single line beginning with \c{\\e}, indicating the
emphasis in that line. The emphasis line contains the letters \c{b}
and \c{i} (for \q{bold} and \q{italic}, although some output formats
might render \c{i} as underlining instead of italics), positioned
to line up under the parts of the text that you want emphasised.

For example, if you wanted to do syntax highlighting on the above C
code by highlighting the preprocessor command in italic and the
keywords in bold, you might do it like this:

\c \c #include <stdio.h>
\c \e iiiiiiiiiiiiiiiiii
\c \c
\c \c int main(int argc, char **argv) {
\c \e bbb      bbb       bbbb
\c \c     printf("hello, world\n");
\c \c     return 0;
\c \e     bbbbbb
\c \c }

and Halibut would generate:

\quote{

\c #include <stdio.h>
\e iiiiiiiiiiiiiiiiii
\c
\c int main(int argc, char **argv) {
\e bbb      bbb       bbbb
\c     printf("hello, world\n");
\c     return 0;
\e     bbbbbb
\c }

}

Note that not every \c{\\c} line has to be followed by a \c{\\e}
line; they're optional.

Also, note that highlighting within a code paragraph is
\e{discretionary}. Not all of Halibut's output formats can support
it (plain text, in particular, has no sensible way to do it). Unless
you know you are using a restricted range of output formats, you
should use highlighting in code paragraphs \e{only} as a visual aid,
and not rely on it to convey any vital semantic content.

\S{input-lists} \c{\\b}, \c{\\n}, \c{\\dt}, \c{\\dd}, \c{\\lcont}: Lists

Halibut supports bulletted lists, numbered lists and description
lists.

\S2{input-list-bullet} \c{\\b}: Bulletted lists

To create a bulletted list, you simply prefix each paragraph
describing a bullet point with the command \c{\\b}. For example, this
Halibut input:

\c Here's a list:
\c
\c \b One.
\c
\c \b Two.
\c
\c \b Three.

would produce this Halibut output:

\quote{
Here's a list:

\b One.

\b Two.

\b Three.
}

\S2{input-list-number} \c{\\n}: Numbered lists

Numbered lists are just as simple: instead of \c{\\b}, you use
\c{\\n}, and Halibut takes care of getting the numbering right for
you. For example:

\c Here's a list:
\c
\c \n One.
\c
\c \n Two.
\c
\c \n Three.

This produces the Halibut output:

\quote{
Here's a list:

\n One.

\n Two.

\n Three.
}

The disadvantage of having Halibut sort out the list numbering for
you is that if you need to refer to a list item by its number, you
can't reliably do so. To get round this, Halibut allows an optional
keyword in braces after the \c{\\n} command. This keyword can then
be referenced using the \c{\\k} or \c{\\K} command (see
\k{input-xref}) to provide the number of the list item. For example:

\c Here's a list:
\c
\c \n One.
\c
\c \n{this-one} Two.
\c
\c \n Three.
\c
\c \n Now go back to step \k{this-one}.

This produces the following output:

\quote{
Here's a list:

\n One.

\n{this-one} Two.

\n Three.

\n Now go back to step \k{this-one}.
}

\S2{input-list-description} \c{\\dt} and \c{\\dd}: Description lists

To wrie a description list, you prefix alternate paragraphs with the
\c{\\dt} (\q{described thing}) and \c{\\dd} (description) commands.
For example:

\c \dt Pelican
\c
\c \dd This is a large bird with a big beak.
\c
\c \dt Panda
\c
\c \dd This isn't.

This produces the following output:

\quote{

\dt Pelican

\dd This is a large bird with a big beak.

\dt Panda

\dd This isn't.

}

\S2{input-list-continuation} Continuing list items into further
paragraphs

All three of the above list types assume that each list item is a
single paragraph. For a short, snappy list in which each item is
likely to be only one or two words, this is perfectly sufficient;
but occasionally you will find you want to include several
paragraphs in a single list item, or even to nest other types of
paragraph (such as code paragraphs, or other lists) inside a list
item.

To do this, you use the \c{\\lcont} command. This is a command which
can span \e{multiple} paragraphs.

After the first paragraph of a list item, include the text
\c{\\lcont\{}. This indicates that the subsequent paragraph(s) are a
\e{continuation} of the list item that has just been seen. So you
can include further paragraphs, and eventually include a closing
brace \c{\}} to finish the list continuation. After that, you can
either continue adding other items to the original list, or stop
immediately and return to writing normal paragraphs of text.

Here's a (long) example.

\c Here's a list:
\c
\c \n One. This item is followed by a code paragraph:
\c
\c \lcont{
\c
\c \c code
\c \c paragraph
\c
\c }
\c
\c \n Two. Now when I say \q{two}, I mean:
\c
\c \lcont{
\c
\c \n Two, part one.
\c
\c \n Two, part two.
\c
\c \n Two, part three.
\c
\c }
\c
\c \n Three.

The output produced by this fragment is:

\quote{

Here's a list:

\n One. This item is followed by a code paragraph:

\lcont{

\c code
\c paragraph

}

\n Two. Now when I say \q{two}, I mean:

\lcont{

\n Two, part one.

\n Two, part two.

\n Two, part three.

}

\n Three.

}

This syntax seems a little bit inconvenient, and perhaps
counter-intuitive: you might expect the enclosing braces to have to
go around the \e{whole} list item, rather than everything except the
first paragraph.

\c{\\lcont} is a recent addition to the Halibut input language;
previously, \e{all} lists were required to use no more than one
paragraph per list item. So it's certainly true that this feature
looks like an afterthought because it \e{is} an afterthought, and
it's possible that if I'd been designing the language from scratch
with multiple-paragraph list items in mind, I would have made it
look different.

However, the advantage of doing it this way is that no enclosing
braces are required in the \e{common} case: simple lists with only
one paragraph per item are really, really easy to write. So I'm not
too unhappy with the way it turned out; it obeys the doctrine of
making simple things simple, and difficult things possible.

Note that \c{\\lcont} can only be used on \c{\\b}, \c{\\n} and
\c{\\dd} paragraphs; it cannot be used on \c{\\dt}.

\S{input-rule} \c{\\rule}: Horizontal rules

The command \c{\\rule}, appearing on its own as a paragraph, will
cause a horizontal rule to be drawn, like this:

\c Some text.
\c
\c \rule
\c
\c Some more text.

This produces the following output:

\quote{

Some text.

\rule

Some more text.

}

\S{input-quote} \c{\\quote}: Indenting multiple paragraphs as a long
quotation

Quoting verbatim text using a code paragraph (\k{input-codepara}) is
not always sufficient for your quoting needs. Sometimes you need to
quote some normally formatted text, possibly in multiple paragraphs.

To do this, you can use the \c{\\quote} command. Like \c{\\lcont},
this is a command which expects to enclose at least one paragraph
and possibly more. Simply write \c{\\quote\{} at the beginning of
your quoted section, and \c{\}} at the end, and the paragraphs in
between will be formatted to indicate that they are a quotation.

(This very manual, in fact, uses this feature a lot: all of the
examples of Halibut source followed by Halibut output have the
output quoted using \c{\\quote}.)

Here's some example Halibut input:

\c In \q{Through the Looking Glass}, Lewis Carroll wrote:
\c
\c \quote{
\c
\c \q{The question is,} said Alice, \q{whether you \e{can} make
\c words mean so many different things.}
\c
\c \q{The question is,} said Humpty Dumpty, \q{who is to be master -
\c that's all.}
\c
\c }
\c
\c So now you know.

The output generated by this is:

\quote{

In \q{Through the Looking Glass}, Lewis Carroll wrote:

\quote{

\q{The question is,} said Alice, \q{whether you \e{can} make
words mean so many different things.}

\q{The question is,} said Humpty Dumpty, \q{who is to be master -
that's all.}

}

So now you know.

}

\S{input-sections} \c{\\C}, \c{\\H}, \c{\\S}, \c{\\A}, \c{\\U}:
Chapter and section headings

\K{intro-features} mentions that Halibut numbers the sections of
your document automatically, and can generate cross-references to
them on request; \k{input-xref} describes the \c{\\k} and \c{\\K}
commands used to generate the cross-references. This section
describes the commands used to set up the sections in the first
place.

A paragraph beginning with the \c{\\C} command defines a chapter
heading. The \c{\\C} command expects to be followed by a pair of
braces containing a keyword for the chapter; this keyword can then
be used with the \c{\\k} and \c{\\K} commands to generate
cross-references to the chapter. After the closing brace, the rest
of the paragraph is used as the displayed chapter title. So the
heading for the current chapter of this manual, for example, is
written as

\c \C{input} Halibut input format

and this allows me to use the command \c{\\k\{input\}} to generate a
cross-reference to that chapter somewhere else.

The next level down from \c{\\C} is \c{\\H}, for \q{heading}. This
is used in exactly the same way as \c{\\C}, but section headings
defined with \c{\\H} are considered to be part of a containing
chapter, and will be numbered with a pair of numbers. After \c{\\H}
comes \c{\\S}, and if necessary you can then move on to \c{\\S2},
\c{\\S3} and so on.

For example, here's a sequence of heading commands. Normally these
commands would be separated at least by blank lines (because each is
a separate paragraph), and probably also by body text; but for the
sake of brevity, both of those have been left out in this example.

\c \C{foo} Using Foo
\c \H{foo-intro} Introduction to Foo
\c \H{foo-running} Running the Foo program
\c \S{foo-inter} Running Foo interactively
\c \S{foo-batch} Running Foo in batch mode
\c \H{foo-trouble} Troubleshooting Foo
\c \C{bar} Using Bar instead of Foo

This would define two chapters with keywords \c{foo} and \c{bar},
which would end up being called Chapter 1 and Chapter 2 (unless
there were other chapters before them). The sections \c{foo-intro},
\c{foo-running} and \c{foo-trouble} would be referred to as Section
1.1, Section 1.2 and Section 1.3 respectively; the subsections
\c{foo-inter} and \c{foo-batch} would be Section 1.2.1 and Section
1.2.2. If there had been a \c{\\S2} command within one of those, it
would have been something like Section 1.2.1.1.

If you don't like the switch from \c{\\H} to \c{\\S}, you can use
\c{\\S1} as a synonym for \c{\\S} and \c{\\S0} as a synonym for
\c{\\H}. Chapters are still designated with \c{\\C}, because they
need to be distinguished from other types of chapter such as
appendices. (Personally, I like the \c{\\C},\c{\\H},\c{\\S} notation
because it encourages me to think of my document as a hard disk :-)

You can define an appendix by using \c{\\A} in place of \c{\\C}.
This is no different from a chapter except that it's given a letter
instead of a number, and cross-references to it will say \q{Appendix
A} instead of \q{Chapter 9}. Subsections of an appendix will be
numbered \q{A.1}, \q{A.2}, \q{A.2.1} and so on.

If you want a particular section to be referred to as something
other than a \q{chapter}, \q{section} or \q{appendix}, you can
include a second pair of braces after the keyword. For example, if
you're writing a FAQ chapter and you want cross-references between
questions to refer to \q{question 1.2.3} instead of \q{section
1.2.3}, you can write each section heading as

\c \S{question-about-fish}{Question} What about fish?

(The word \q{Question} should be given with an initial capital
letter. Halibut will lower-case it when you refer to it using
\c{\\k}, and will leave it alone if you use \c{\\K}.)

This technique allows you to change the designation of
\e{particular} sections. To make an overall change in what \e{every}
section is called, see \k{input-config}.

Finally, the \c{\\U} command defines an \e{unnumbered} chapter.
These sometimes occur in books, for specialist purposes such as
\q{Bibliography} or \q{Acknowledgements}. \c{\\U} does not expect a
keyword argument, because there is no sensible way to generate an
automatic cross-reference to such a chapter anyway.

\S{input-blurb} \c{\\preamble}, \c{\\copyright}, \c{\\title},
\c{\\versionid}: Miscellaneous blurb commands

These four commands define a variety of special paragraph types.
They are all used in the same way: you put the command at the start
of a paragraph, and then just follow it with normal text, like this:

\c \title My First Manual

The four special paragraph types are:

\dt \c{\\title}

\dd This defines the overall title of the entire document. This
title is treated specially in some output formats (for example, it's
used in a \cw{<title>} tag in the HTML output), so it needs a
special paragraph type to point it out.

\dt \c{\\preamble}

\dd This command indicates ordinary text that appears \e{before} the
first chapter heading. \#{FIXME: figure out what Halibut actually
does do with ordinary text before a chapter heading. Then figure out
whether that's sensible. Then figure out what \preamble has to offer
that's perceptibly different. Fix if necessary.}

\dt \c{\\copyright}

\dd This command indicates that the paragraph attached to it
contains a copyright statement for the document. This text is
usually displayed inline, in the same way as \c{\\preamble} text,
but in some output formats it is given additional special treatment.
For example, Windows Help files have a standard slot in which to
store a copyright notice, so that other software can display it
prominently.

\dt \c{\\versionid}

\dd This command indicates that the paragraph contains a version
identifier, such as those produced by CVS (of the form \c{$\#{hope this
defuses CVS}Id: thingy.but,v 1.6 2004/01/01 16:47:48 simon Exp $}).
This text will be tucked away somewhere unobtrusive, so that anyone
wanting to (for example) report errors to the document's author can
pick out the version IDs and send them as part of the report, so
that the author can tell at a glance which revision of the document
is being discussed.

\S{input-commentpara} \c{\\#}: Whole-paragraph comments

\K{input-inline-comment} describes the use of the \c{\\#} command to
put a short comment in the middle of a paragraph.

If you need to use a \e{long} comment, Halibut also allows you to
use \c{\\#} without braces, to indicate that an entire paragraph is
a comment, like this:

\c Here's a (fairly short) paragraph which will be displayed.
\c
\c \# Here's a comment paragraph which will not be displayed, no
\c matter how long it goes on. All I needed to indicate this was the
\c single \# at the start of the paragraph; I don't need one on
\c every line or anything like that.
\c
\c Here's another displayed paragraph.

When run through Halibut, this produces the following output:

\quote{

Here's a (fairly short) paragraph which will be displayed.

\# Here's a comment paragraph which will not be displayed, no
matter how long it goes on. All I needed to indicate this was the
single \# at the start of the paragraph; I don't need one on
every line or anything like that.

Here's another displayed paragraph.

}

\H{input-biblio} Creating a bibliography

If you need your document to refer to other documents (research
papers, books, websites, whatever), you might find a bibliography
feature useful.

You can define a bibliography entry using the \c{\\B} command. This
looks very like the \c{\\C} command and friends: it expects a
keyword in braces, followed by some text describing the document
being referred to. For example:

\c \B{freds-book} \q{The Taming Of The Mongoose}, by Fred Bloggs.
\c Published by Paperjam & Notoner, 1993.

If this bibliography entry appears in the finished document, it will
look something like this:

\quote{

\B{freds-book} \q{The Taming Of The Mongoose}, by Fred Bloggs.
Published by Paperjam & Notoner, 1993.

}

I say \q{if} above because not all bibliography entries defined
using the \c{\\B} command will necessarily appear in the finished
document. They only appear if they are referred to by a \c{\\k}
command (see \k{input-xref}). This allows you to (for example)
maintain a single Halibut source file with a centralised database of
\e{all} the references you have ever needed in any of your writings,
include that file in every document you feed to Halibut, and have it
only produce the bibliography entries you actually need for each
particular document. (In fact, you might even want this centralised
source file to be created automatically by, say, a Perl script from
BibTeX input, so that you can share the same bibliography with users
of other formatting software.)

If you really want a bibliography entry to appear in the document
even though no text explicitly refers to it, you can do that using
the \c{\\nocite} command:

\c \nocite{freds-book}

Normally, each bibliography entry will be referred to (in citations
and in the bibliography itself) by a simple reference number, such
as \k{freds-book}. If you would rather use an alternative reference
notation, such as [Fred1993], you can use the \c{\\BR}
(\q{Bibliography Rewrite}) command to specify your own reference
format for a particular book:

\c \BR{freds-book} [Fred1993]

\H{input-index} Creating an index

Halibut contains a comprehensive indexing mechanism, which attempts
to be reasonably easy to use in the common case in spite of its
power.

\S{input-index-simple} Simple indexing

In normal usage, you should be able to add index terms to your
document simply by using the \c{\\i} command to wrap one or two
words at a time. For example, if you write

\c The \i{hippopotamus} is a particularly large animal.

then the index will contain an entry under \q{hippopotamus},
pointing to that sentence (or as close to that sentence as the
output format sensibly permits).

You can wrap more than one word in \c{\\i} as well:

\c We recommend using a \i{torque wrench} for this job.

\S{input-index-special} Special cases of indexing

If you need to index a computer-related term, you can use the
special case \c{\\i\\c}:

\c The \i\c{grep} command is what you want here.

This will cause the word \q{grep} to appear in code style, as if the
\c{\\i} were not present and the input just said \c{\\c\{grep\}};
the word will also appear in code style in the actual index.

If you want to simultaneously index and emphasise a word, there's
another special case \c{\\i\\e}:

\c This is what we call a \i\e{paper jam}.

This will cause the words \q{paper jam} to be emphasised in the
document, but (unlike the behaviour of \c{\\i\\c}) they will \e{not}
be emphasised in the index. This different behaviour is based on an
expectation that most people indexing a word of computer code will
still want it to look like code in the index, whereas most people
indexing an emphasised word will \e{not} want it emphasised in the
index.

Sometimes you might want to index a term which is not explicitly
mentioned, but which is highly relevant to the text and you think
that somebody looking up that term in the index might find it useful
to be directed here. To do this you can use the \c{\\I} command, to
create an \e{invisible} index tag:

\c If your printer runs out of toner, \I{replacing toner
\c cartridge}here is what to do:

This input will produce only the output \q{If your printer runs out
of toner, here is what to do}; but an index entry will show up under
\q{replacing toner cartridge}, so that if a user thinks the obvious
place to start in the index is under R for \q{replacing}, they will
find their way here with a minimum of fuss.

(It's worth noting that there is no functional difference between
\c{\\i\{foo\}} and \c{\\I\{foo\}foo}. The simple \c{\\i} case is
only a shorthand for the latter.)

Finally, if you want to index a word at the start of a sentence, you
might very well not want it to show up with a capital letter in the
index. For this, Halibut provides the \c{\\ii} command, for \q{index
(case-)insensitively}. You use it like this:

\c \ii{Lions} are at the top of the food chain in this area.

This is equivalent to \c{\\I\{lions\}Lions}; in other words, the
text will say \q{Lions}, but it will show up in the index as
\q{lions}. The text inside \c{\\ii} is converted entirely into lower
case before being added to the index data.

\S{input-index-rewrite} Fine-tuning the index

Halibut's index mechanism as described so far still has a few
problems left:

\b In a reasonably large index, it's often difficult to predict
which of several words a user will think of first when trying to
look something up. For example, if they want to know how to replace
a toner cartridge, they might look up \q{replacing} or they might
look up \q{toner cartridge}. You probably don't really want to have
to try to figure out which of those is more likely; instead, what
you'd like is to be able to effortlessly index the same set of
document locations under \e{both} terms.

\b Also, you may find you've indexed the same concept under multiple
different index terms; for example, there might be several instances
of \c{\\i\{frog\}} and several of \c{\\i\{frogs\}}, so that you'd
end up with two separate index entries for what really ought to be
the same concept.

\b You might well not want the word \q{\cw{grep}} to appear in the
index without explanation; you might prefer it to say something more
verbose such as \q{\cw{grep} command}, so that a user encountering
it in the index has some idea of what it is \e{without} having to
follow up the reference. However, you certainly don't want to have
to write \c{\\I\{\\cw\{grep\} command\}\\c\{grep\}} every time you
want to add an index term for this! You wanted to write
\c{\\i\\c\{grep\}} as shown in the previous section, and tidy it all
up afterwards.

All of these problems can be cleaned up by the \c{\\IM} (for
\q{Index Modification}) command. \c{\\IM} expects to be followed by
one or more pairs of braces containing index terms as seen in the
document, and then a piece of text (not in braces) describing how it
should be shown in the index.

So to rewrite the \c{grep} example above, you might do this:

\c \IM{grep} \cw{grep} command

This will arrange that the set of places in the document where you
asked Halibut to index \q{\cw{grep}} will be listed under
\q{\cw{grep} command} rather than just under \q{\cw{grep}}.

You can specify more than one index term in a \c{\\IM} command; so
to merge the index terms \q{frog} and \q{frogs} into a single term,
you might do this:

\c \IM{frog}{frogs} frog

This will arrange that the single index entry \q{frog} will list
\e{all} the places in the document where you asked Halibut to index
either \q{frog} or \q{frogs}.

You can use multiple \c{\\IM} commands to replicate the same set of
document locations in more than one index entry. For example:

\c \IM{replacing toner cartridge} replacing toner cartridge
\c \IM{replacing toner cartridge} toner cartridge, replacing

This will arrange that every place in the document where you have
indexed \q{replacing toner cartridge} will be listed both there
\e{and} under \q{toner cartridge, replacing}, so that no matter
whether the user looks under R or under T they will stil find their
way to the same parts of the document.

In this example, note that although the first \c{\\IM} command
\e{looks} as if it's a tautology, it is still necessary, because
otherwise those document locations will \e{only} be indexed under
\q{toner cartridge, replacing}. If you have \e{no} explicit \c{\\IM}
commands for a particular index term, then Halibut will assume a
default one (typically \c{\\IM\{foo\}\_foo}, although it might be
\c{\\IM\{foo\}\_\\c\{foo\}} if you originally indexed using
\c{\\i\\c}); but as soon as you specify an explicit \c{\\IM},
Halibut discards its default implicit one, and you must then specify
that one explicitly as well if you wanted to keep it.

\H{input-config} Configuring Halibut

Halibut uses the \c{\\cfg} command to allow you to configure various
aspects of its functionality.

The \c{\\cfg} command expects to be followed by at least one pair of
braces, and usually more after that. The first pair of braces
contains a keyword indicating what aspect of Halibut you want to
configure, and the meaning of the one(s) after that depends on the
first keyword.

The current list of configuration keywords in the main Halibut code
is quite small. Here it is in full:

\dt \cw{\\cfg\{chapter\}\{}\e{new chapter name}\cw{\}}

\dd This tells Halibut that you don't want to call a chapter a
chapter any more. For example, if you give the command
\cw{\\cfg\{chapter\}\{Book\}}, then any chapter defined with the
\c{\\C} command will be labelled \q{Book} rather than \q{Chapter},
both in the section headings and in cross-references. This is
probably most useful if your document is not written in English.

\dt \cw{\\cfg\{section\}\{}\e{new section name}\cw{\}}

\dd Exactly like \c{chapter}, but changes the name given to
subsections of a chapter.

\dt \cw{\\cfg\{appendix\}\{}\e{new appendix name}\cw{\}}

\dd Exactly like \c{appendix}, but changes the name given to
appendices.

In addition to these configuration commands, there are also
configuration commands provided by each individual output format.
These configuration commands are discussed along with each output
format, in FIXME.

\H{input-macro} Defining macros

If there's a complicated piece of Halibut source which you think
you're going to use a lot, you can define your own Halibut command
to produce that piece of source.

In \k{input-unicode}, there is a sample piece of code which prints a
Euro sign, or replaces it with \q{EUR} if the Euro sign is not
available:

\c This is likely to cost \u20AC{EUR }2500 at least.

If your document quotes a \e{lot} of prices in Euros, you might not
want to spend all your time typing that out. So you could define a
macro, using the \c{\\define} command:

\c \define{eur} \u20AC{EUR }

Then you can just write

\c This is likely to cost \eur 2500 at least.

\# FIXME: need a NOP to allow text to directly follow this. \.?