Java SE 24 & JDK 24
DRAFT 24-loom+10-110

JavaDoc Documentation Comment Specification for the Standard Doclet (JDK 24)

This document specifies the form of documentation comments recognized by the standard doclet for the javadoc tool in JDK 24, used to generate HTML documentation for an API.

In the context of the javadoc tool, the interpretation of the content of a documentation comment is up to doclet that is used to process the comment. Other doclets may accept the same syntax as the standard doclet, or they may support an alternate syntax. However, due to support in many tools, the syntax supported by the standard doclet has become a de facto standard.

Documentation Comments

Documentation comments are stylized comments appearing in source code, near to the declarations that they serve to document.

Documentation comments are recognized only when placed immediately before the declaration of a module, package, class, interface, constructor, method, annotation interface element, enum member, or field. In other words, a documentation comment should appear before any annotation, modifier, keyword, identifier, or other Java construct that is part of the declaration.

There are two kinds of documentation comments: traditional documentation comments, and Markdown documentation comments.

Either kind, or both kinds, may be used in any one source file.

Only one documentation comment per declaration statement is recognized. If there are multiple documentation comments before a declaration, the one closest to the beginning of the declaration is used.

Documentation may also be provided in certain files:

HTML files are identified by a filename ending in .html. The content of the <main> element of an HTML file, or the <body> element if there is no <main> element is processed as if it were the content of a traditional documentation comment.

Markdown files are identified by a filename ending in .md. The content of a Markdown file is processed as if it were the content of a Markdown documentation comment.

Traditional Documentation Comments

Traditional documentation comments are traditional comments that begin with /**. If any line in such a comment begins with asterisks after any leading whitespace, the leading whitespace and asterisks are removed. Any whitespace appearing after the asterisks is not removed.

A traditional documentation comment may contain HTML content, inline tags, and block tags. You can also use escape sequences to represent characters that would otherwise be inconvenient or difficult to represent.

Markdown Documentation Comments

Documentation comments containing Markdown are composed of a series of consecutive lines, each beginning with optional whitespace followed by ///.

A Markdown documentation comment may contain Markdown content (including HTML where necessary), inline tags, and block tags.

General Syntax

The overall form of a documentation comment is an initial main description, followed by a series of block tags, which provide additional information about the declaration to which the comment applies. Descriptive text may include inline tags, HTML content, and Markdown content, as described below.

It is possible to have a comment with only block tags and no main description.

For historical reasons, the documentation comment for a package may instead be provided in a file called package.html in a source directory for the package. In this case, the documentation comment is the content of the <body> tag, and all references to Java types (for example, in see tags) must be fully qualified. The standard doclet also allows additional documentation to be provided in files such as overview.html. The rules for such content are the same as for package.html.

Main Description

The main description in a documentation comment is the content from the beginning of the comment, up to the first block tag, if there are any, or to the end of the comment, if there are none. Leading and trailing whitespace is ignored. If there is no such content, the main description is said to be missing.

The main description cannot continue after any block tags.

Examples of a missing main description.

The first sentence of the main description should be a summary sentence that contains a concise but complete description of the declared entity.

Block Tags

Block tags are of the form @identifier content and give additional details to be incorporated into the generated documentation. Each block tag must appear at the beginning of a line, ignoring leading asterisks, whitespace characters, and the initial comment delimiter (/**). This means you can use the @ character elsewhere in the text, and it will not be interpreted as the start of a tag. If you want to start a line with the @ character and not have it be interpreted, then use the HTML entity &#064;. The content of a block tag is any text following the tag up to, but not including, either the next block tag, or the end of the documentation comment. The content can span multiple lines.

There can be any number of block tags; some types of tags can be repeated while others cannot.

Inline Tags

Inline tags are of the form {@identifier content} and provide details within the context of the enclosing description. They may generally appear wherever descriptive text and HTML is permitted, although some inline tags may only be used at the beginning of the main description.

Some inline tags may contain free-form text. When such text explicitly contains braces, the braces must be "balanced", implying an equal number of appropriately nested left brace and right brace characters, so that the closing brace of the inline tag can be determined. No other lexical analysis of the text is performed; in particular, there is no special consideration of characters like ', ", \, and @.

Lines beginning with @ that are enclosed within an inline tag are not considered as beginning a block tag.

When the text content is HTML, it may be possible to use entities &lbrace; and &rbrace; to represent unbalanced braces.

HTML Content

HTML content is not formally checked, although some tools may provide some amount of checking to help catch common errors.

In order to be able to generate documentation that conforms to the appropriate standards, the following considerations should be taken into account when using HTML constructs in a documentation comment:

Markdown Content

Because horizontal whitespace at the beginning and end of each line of Markdown code may be significant, the content of such a comment is determined as follows:

Note: The policy to remove leading incidental whitespace is similar to that for String.stripIndent, except that there is no need for any special treatment for a trailing blank line.

There are no restrictions on the characters that may appear after the /// on each line of the comment. In particular, the comment may contain code samples which may contain comments of their own:

/// Here is an example:
///
/// ```
/// /** Hello World! */
/// public class HelloWorld {
///     public static void main(String... args) {
///         System.out.println("Hello World!"); // the traditional example
///     }
/// }
/// ```

As well as serving to visually distinguish the new kind of documentation comment, the use of end-of-line (//) comments eliminates the restrictions on the content of the comment that are inherent with the use of traditional (/* ... */) comments. In particular, it is not possible to use the character sequence */ within a traditional comment, (JLS 3.7) although it may be desirable to do so when writing example code containing traditional comments, strings containing "glob" expressions, and strings containing regular expressions.

Note: For a blank line to be included in the comment, it must begin with any optional whitespace followed by ///. A completely blank line will cause any preceding and following comment to be treated as separate comments, in which case, all but the last comment will be discarded, and only the last comment will be considered as a documentation comment for any declaration that may follow.

For example:

/// This is an example ...
///
/// ... of a 3-line comment containing a blank line.

The blank line in the following example will terminate the first comment, so that the comment that comes after the blank line will be treated as a separate comment.

/// This comment will be treated as a "dangling comment" and will be ignored.

/// This is the comment for the following declaration.
public void m() { }

The same is true for any other comment not beginning with /// that may appear between two /// comments.

Syntax

The overall syntax of a Markdown documentation comment is that of the CommonMark variant of Markdown. Enhancements to links allow convenient linking to other program elements. Simple GFM pipe tables are supported, as are all JavaDoc tags.

Headings

Both ATX headings and setext headings are supported. Unlike HTML headings, Markdown headings should start at level 1 and increase from there, whatever the context of the documentation comment. (They will be translated to the correct heading level in the generated documentation.)

The general form of a reference link is [text][label] where text is the text to be displayed, and label is a label for the destination of the link, and which is subsequently defined by a line of the form [label]: uri. Collapsed and shortcut forms can also be used when the label is the same as the text.

In a documentation comment, if the label for a reference link is to be defined explicitly, it must be defined in the same documentation comment. But, if it is not defined and if it syntactically matches that of a reference to a program element, then it is implicitly defined as a reference to that program element. It is an error if the label looks like a reference to a program element but there is no such element.

The following all represent links to java.lang.Object#hashCode(), albeit with different text to be displayed:

You can link to any kind of program element, as shown in the following examples:

/// * a module [java.base/]
/// * a package [java.util]
/// * a class [String]
/// * a field [String#CASE_INSENSITIVE_ORDER]
/// * a method [String#chars()]

Inline links are also supported in documentation comments, but the helpful recognition of program elements as labels means that reference links are generally more useful than inline links.

According to the standard rules for reference links, you must escape any use of square brackets within a reference. This might occur when you are creating a reference to a method with an array parameter. The following shows a link to String.copyValueOf(char[])

The normal rendering for a reference link is to use the default plain font. If you wish to use a monospace font, you must enclose the text within a code span, using backticks. For references to program elements, monospace font will be used automatically if the text is the same as the reference (such as when using the collapsed or shortcut forms), and the default plain font otherwise. The following shows how these forms correspond to the existing link and linkplain tags.

Note: You cannot use a reference link in the first sentence of any documentation comment that refers to a user-defined link reference definition elsewhere in the documentation comment; you can use an inline link in this situation instead.

Tables

Simple tables are supported, using the syntax defined in GitHub Flavored Markdown. For example, a simple table might be written as follows:

/// | Latin | Greek |
/// |-------|-------|
/// | a     | alpha |
/// | b     | beta  |
/// | c     | gamma |

Note: Captions and other features that may be required for accessibility are not supported. In such situations, the use of HTML tables is still recommended.

Inline Tags and Block Tags

Both inline tags and block tags may be used in Markdown documentation comments, although neither may be used within the literal text of code spans and fenced or indented code blocks.

For those tags that may contain text with markup, in a Markdown documentation comment that markup will also be in Markdown format.

As in traditional documentation comments, the content of inline tags may span multiple lines, including blank lines.

When using a reference link with an explicit link label, the reference and the label must appear in the same part of the comment: in either the main description, or in any inline or block tag. A reference link may not refer to a label defined in a different part of the comment, or in a different comment altogether.

The inheritDoc tag is used to include documentation from one or more supertypes. There is no requirement that the format of the comment containing the tag is the same as the format of the comment containing the documentation to be inherited. While it may be the case that comments within a library may all use the same format, that may be less likely when handling comments in different libraries, written by different authors.

Syntax Highlighting and Embedded Languages

The opening fence in a fenced code block may be followed by an info string, the first word of which is used to derive the CSS class name in the corresponding generated HTML, and which may be used by JavaScript libraries to enable syntax highlighting (such as with Prism) or rendering diagrams (such as with Mermaid).

Note: You can add JavaScript libraries to your documentation by using the javadoc --add-script option.

Escape Sequences

The following escape sequences are supported in traditional documentation comments wherever text, entities and HTML may appear, to represent characters that would otherwise be inconvenient or difficult to represent:

Escape sequences are context-sensitive, and can only be used where the use of the escaped character by itself would have a different syntactic interpretation. In other situations, these character sequences are taken literally, without additional interpretation.

Escape sequences cannot be used in inline tags that contain literal text; this includes code, literal, snippet, and user-defined tags.

References

References are the constructs in a documentation comment that refer to elements in the surrounding declarations. Depending on the context, they may refer to modules, packages, classes and interfaces, constructors, methods, annotation members, fields, enum members, parameters, record components and the names of exceptions that may be thrown by a method or constructor.

The most general form of a reference is:

This form is used by the see, link and linkplain tags.

Leading components can be omitted when they can be inferred from the surrounding context. Trailing components can be omitted when they are not required. Generally, the reference is evaluated in the scope in which the documentation comment exists. In particular, import statements for the compilation unit are taken into account when evaluating class and interface names.

The class may be any top-level or nested class or interface.

The member may be any constructor, method, annotation member, field or enum member, but not a nested class or interface. As in Java source code, a constructor is identified by using the name of its class. The name of a constructor or method should normally be followed by the list of its parameter types, enclosed in parentheses, although the parameter types and parentheses can be omitted if the method or constructor is not overloaded and the name is not also that of a field or enum member in the same class or interface. When a parameter list is given, whitespace characters may appear between tokens in the parameter list; whitespace characters may not appear elsewhere in the reference. When the reference is to a member of the same class as that containing the documentation comment, all parts of the reference up to and including the # may be omitted, although the '#' may be retained for clarity.

Parameterized types may be used in the class and member parts of the reference; annotations may not be used anywhere in the reference. Whitespace characters may occur between tokens within the parameter list for a constructor or method. A trailing / can be added to a name to refer to a module in the presence of a package or class with the same name.

Note: You cannot refer to the declaration of a specific parameter or record component with this form.

An alternative form is provided to generate references to arbitrary URI fragments in the generated documentation such as headings in documentation comments. This form uses a double hash mark (##) as separator:

fragment is interpreted as URI fragment within the page documenting the specified program element.

Other tags, such as param, throws, and serialField may only provide support for the specific kinds of references that are relevant to each tag. See the description of individual tags for more details.

Method Documentation

A documentation comment for a method declaration must at least provide the following:

It is an error if any item described in that list is missing from a documentation comment of a method declaration and either of following is true:

Otherwise, a missing item is considered as if it were provided with {@inheritDoc} as its content.

For the purpose of this specification an overriding method declaration is a method declaration that could be annotated with @Override (JLS 9.6.4.4), but is not an accessor method declaration for a record component.

Method documentation allows inheritance by omission: if a documentation comment for a method declaration provides only some of the items, the rest are assumed to be inherited. For example, if this is provided:

/**
 * @param scale a non-zero number
 * @throws IllegalArgumentException if scale is 0
 */
 @Override
 <T> T magnify(int scale, T element) throws MagnificationException

This is assumed:

/**
 * {@inheritDoc}
 *
 * @param <T> {@inheritDoc}
 * @param scale a non-zero number
 * @param element {@inheritDoc}
 * @return {@inheritDoc}
 * @throws IllegalArgumentException if scale is 0
 * @throws MagnificationException {@inheritDoc}
 */
 @Override
 <T> T magnify(int scale, T element) throws MagnificationException

However, not all missing items can be inherited. For example, if a method declares an exception X that is not documented by any of the methods that this method overrides, then documentation item for that exception cannot be inherited and can only be provided by this method's documentation comment. (Adding or assuming @throws X {@inheritDoc} would be an error.) If that item is not provided, it is considered missing.

(@Override is only added to these examples to emphasize that the declarations are overriding; that annotation has no effect on documentation.)

Note: A documentation comment consisting of a sole inheritDoc tag explicitly provides only the main description of a method:

/**
 * {@inheritDoc}
 */

Other items, if any, are inherited by omission.

Note: It is the overriding method, not the overridden method, that determines which items are missing and which are not.

For example, if an overriding method declares a thrown exception X, but that method's documentation comment does not contain @throws X ..., then that documentation item is missing. If an overriding method does NOT declare a thrown exception, even if the overridden method does, then that documentation item is NOT missing.

This behavior is usually noticed when dealing with unchecked exceptions. Since unchecked exceptions do not normally appear in the throws clause, unless explicitly documented, they will not be missing. This gives rise to a misconception that the standard doclet treats checked exceptions differently from unchecked exceptions, which it does not.

In fact, a similar behavior might be replicated with checked exceptions. Consider a method that declares two checked exceptions X and Y, such that X is a supertype of Y. If an overriding method declares X but not Y, then @throws Y ... won't be missing.

Overriding Methods in Classes and Interfaces

When a method declaration overrides a method in the superclass or an extended superinterface, the standard doclet generates the subheading "Overrides" in the documentation for the overriding method.

When a method declaration in a class overrides a method in an interface, the standard doclet generates the subheading "Specified by" in the documentation for the overriding method.

In both cases, a link to the overridden method is included.

Standard Tags

The following sections describe the standard block and inline tags supported by the standard doclet.

Note: The standard doclet also supports user-defined tags conforming to the same general syntactic rules.

author

Adds an "Author" entry with the specified name text to the generated documents when the -author option is used. A documentation comment can contain multiple author tags. You can specify one name per author tag or multiple names per tag. In the former case, the standard doclet inserts a comma (,) and a space between names. In the latter case, the entire text is copied to the generated document without being parsed. Therefore, you can use multiple names per line if you want a localized name separator other than a comma.

If there are no author tags in the description for a member class or interface, the standard doclet will (recursively) look for any such tags in the enclosing class or interface.

Introduced in JDK 1.0.

code

Equivalent to <code>{@literal text }</code>.

Displays text in code font without interpreting the text as HTML markup or nested Javadoc tags. This enables you to use regular angle brackets (< and >) instead of the HTML entities (&lt; and &gt;) in documentation comments, such as in parameter types (<Object>), inequalities (3 < 4), or arrows (->). For example, the documentation comment text {@code A<B>C} displayed in the generated HTML page unchanged as A<B>C. This means that the <B> is not interpreted as bold and is in code font. If you want the same functionality without the code font, then use the literal tag.

Introduced in JDK 1.5.

deprecated

This tag is used in conjunction with the @Deprecated annotation to indicate that this API should no longer be used (even though it may continue to work). The standard doclet moves deprecated text ahead of the main description, placing it in italics and preceding it with a bold warning.

The first sentence of deprecated text should tell the user when the API was deprecated and what to use as a replacement. The standard doclet copies the first sentence to the summary section and index. Subsequent sentences can also explain why it was deprecated. You should include a link tag that points to the replacement API.

Note:

Introduced in JDK 1.0.

docRoot

Represents the relative path to the generated document's (destination) root directory from any generated page. This tag is useful when you want to include a file, such as a copyright page or company logo, that you want to reference from all generated pages. Linking to the copyright page from the bottom of each page is common.

The docRoot tag can be used both on the command line and in a documentation comment. The tag is valid in all documentation comments: overview, module, package, class, interface, constructor, method and field, including the text portion of any tag (such as the return, param and deprecated tags).

For example, on the command line, where the header, footer, or bottom are defined:

javadoc -bottom '<a href="{@docRoot}/copyright.html">Copyright</a>'.

When you use the docRoot tag this way in a make file, some makefile programs require a special way to escape for the brace {} characters. For example, the Inprise MAKE version 5.2 running on Windows requires double braces: {{@docRoot}}. It also requires double (rather than single) quotation marks to enclose arguments to options such as the -bottom option (with the quotation marks around the href argument omitted).

For example, in a documentation comment:

/**
 * See the <a href="{@docRoot}/copyright.html">Copyright</a>.
 */

This tag is needed because the generated documents are in hierarchical directories, as deep as the number of subpackages. The expression <a href="{@docRoot}/copyright.html"> resolves to <a href="../../copyright.html"> for java/lang/Object.java and <a href="../../../copyright.html"> for java/lang/ref/Reference.java.

Introduced in JDK 1.3.

exception

This tag is equivalent to the throws tag, which is now the recommended form.

Introduced in JDK 1.0.

hidden

Hides a program element from the generated API documentation. This tag may be used when it is not otherwise possible to design the API in a way that such items do not appear at all.

Introduced in JDK 9.

index

Declares that a word or phrase, together with an optional short description, should appear in the index files generated by the standard doclet. The index entry will be linked to the word or phrase that will appear at this point in the generated documentation. The description may be used when the word or phrase to be indexed is not clear by itself, such as for an acronym.

Introduced in JDK 9.

inheritDoc

Indicates that a method documentation comment part that contains this tag is the same as that of the overridden method in a superclass or superinterface, which is either determined automatically or specified.

It is an error if the overridden method is missing the corresponding part.

This tag enables you to write more general comments higher up the inheritance hierarchy and to write around the inherited parts.

This tag is valid only in these places in a documentation comment of an overriding declaration of a method:

If the corresponding part of the overridden method's documentation comment itself uses an inheritDoc tag, then that tag is processed first. (This might be applied recursively as required.)

Let m be a method declaration in a class or interface T. If the documentation comment for m contains an inheritDoc tag then the superclass or superinterface (whose overridden method this tag will use the corresponding part of) is determined as follows:

The search for the supertype is performed in two phases: the recursive phase, then the final phase. For simplicity, the search starts from S, where S = T, although there's nothing to be found in T (otherwise, the search would not be required). When starting from S, where S = T, the first condition always breaks into "otherwise" because a method does not override itself (JLS 8.4.8.1).

Recursive phase

If S declares a method that T.m overrides, then S is a supertype, otherwise:

If S is a class that has a direct superclass (JLS 8.1.4) that is not java.lang.Object, then the recursive phase of the search is applied to that superclass. If that application has not found a supertype, then:

If S has direct superinterfaces (JLS 8.1.5), then the recursive phase of the search is applied to each of those superinterfaces, in the order that they are listed in the extends (if S is a class) or implements (if S is an interface) clause, until the supertype found or superinterfaces are exhausted.

Final phase

If the recursive phase of search has not found the supertype, then one final attempt is made to see if java.lang.Object is the supertype. If T is a class and m overrides a method in java.lang.Object, or T is an interface and m's signature is override-equivalent to the signature of a public method in java.lang.Object (JLS 8.4.2), then java.lang.Object is the supertype. Otherwise, the search has not found the supertype.

Note: The search is performed in two phases so as to not prematurely consider java.lang.Object as the supertype. Sometimes java.lang.Object provides overly general documentation for its public methods (such as equals, hashCode, toString) when more specific documentation is available.

For example, consider this schematic hierarchy:

interface A
interface B extends A
interface C
class D implements B, C
interface E
class F extends D implements E
    (A)
     ^
      \
      (B) (C)    [java.lang.Object]
       ^   ^
        \ /
        [D] (E)
         ^   ^
          \ /
          [F]
The same hierarchy represented visually.
The automatic supertype search traverses that hierarchy in the following order: F, D, B, A, C, E, java.lang.Object. 
    (4)
     ^
      \
      (3) (5)    [7]
       ^   ^
        \ /
        [2] (6)
         ^   ^
          \ /
          [1]
The same traversal order represented visually.

Note: The source file for an inherited method must be on the source path for the documentation comment to be available for copy. Neither the class nor its package needs to be passed in on the command line.

Note: When using Markdown comments, there is no requirement that the comment containing the tag and the comment containing the documentation inherited by that tag should either be both Markdown documentation comments or both traditional (not Markdown) comments.

Introduced in JDK 1.4; accepts an optional argument since JDK 22.

Inserts an inline link with a visible text label that points to the documentation for the specified module, package, class, or member name of a referenced class. This tag is valid in all documentation comments: overview, module, package, class, interface, constructor, method and field, including the description part of any tag, such as the return, param and deprecated tags.

This tag is similar to the third form of the see tag. The main difference is that the link tag generates an inline link rather than placing the link in the "See Also" section. The link tag begins and ends with braces to separate it from the rest of the inline text. If you need to use the right brace (}) inside the label, then use the HTML entity notation &#125;.

There is no limit to the number of link tags allowed in a sentence.

For example, here is a comment that refers to the getComponentAt(int, int) method:

Use the {@link #getComponentAt(int, int) getComponentAt} method.

From this code, the standard doclet generates the following HTML (assuming it refers to another class in the same package):

Use the <a href="Component.html#getComponentAt(int,int)">getComponentAt</a> method.

The previous line appears on the web page as:

Use the getComponentAt method.

Introduced in JDK 1.2.

linkplain

Behaves the same as the link tag, except the link label is displayed in plain text rather than code font. Useful when the label is plain text. For example,

Refer to {@linkplain #add() the overridden method}.

is displayed as:

Refer to the overridden method.

Introduced in JDK 1.4.

literal

Displays text without interpreting the text as HTML markup or nested Javadoc tags. This enables you to use angle brackets (< and >) instead of the HTML entities (&lt; and &gt;) in documentation comments, such as in parameter types (<Object>), inequalities (3 < 4), or arrows (->). For example, the documentation comment text {@literal A<B>C} displays unchanged in the generated HTML page in your browser, as A<B>C. The <B> is not interpreted as bold (and it is not in code font). If you want the same functionality with the text in code font, then use the code tag.

Introduced in JDK 1.5.

param

Adds a parameter with the specified parameter name followed by the specified description to the "Parameters" section. The description may continue onto multiple lines. This tag is valid only in a documentation comment for a method, constructor, or class. The parameter name can be the name of a parameter in a method or constructor, or the name of a type parameter of a class, method, or constructor. Use angle brackets (< >) around such a parameter name to indicate the use of a type parameter.

Example of a type parameter of a class:

/**
 * @param <E> Type of element stored in a list
 */
public interface List<E> extends Collection<E> { ... }

Example of parameters, including type parameters, of a method:

/**
 * @param string  the string to be converted
 * @param type    the type to convert the string to
 * @param <T>     the type of the element
 * @param <V>     the value of the element
 */
<T, V extends T> V convert(String string, Class<T> type) { ... }

When an inheritDoc tag is used in the description of a method formal or type parameter, the corresponding part of the overridden method documentation is matched using the parameter's position, not the parameter's name (JLS 8.4.2 and JLS 8.4.4).

For example, consider the following method declaration:

/**
 ...
 * @param scale a non-zero number
 ...
 * @param <T> the type of the element being magnified
 ...
 */
 <T> T magnify(int scale, T element) throws MagnificationException

Then as far as the documentation inheritance is concerned, these two comments are equivalent:

/**
 ...
 * @param s {@inheritDoc}
 * @param e {@inheritDoc}
 ...
 * @param <E> {@inheritDoc}
 */
 @Override
 <E> E magnify(int s, E e) throws MagnificationException

/**
 ...
 * @param scale {@inheritDoc}
 * @param element {@inheritDoc}
 ...
 * @param <T> {@inheritDoc}
 */
 @Override
 <T> T magnify(int scale, T element) throws MagnificationException

If parameter names differ, care should be taken when inheriting such parameter documentation because that parameter name might be referred to from elsewhere both in the overridden and overriding methods.

Introduced in JDK 1.0.

provides

This tag may only appear in the documentation comment for a module declaration, and serves to document an implementation of a service provided by the module. The description may be used to specify how to obtain an instance of this service provider, and any important characteristics of the provider.

Introduced in JDK 9.

return

As a block tag, adds a "Returns" section with the given description, providing details about the values that may be returned by the method.

As an inline tag, provides content for the first sentence of a method's main description, and a "Returns" section, as if "@return description" were also present. In the default English locale, the first sentence is "Returns description .".

This tag is valid only in a documentation comment for a method. As an inline tag, it may only occur at the beginning of a method's main description.

When inheriting documentation, either form of a return tag can be used:

/**
 * {@return {@inheritDoc}}
 ...
 */
<T> T magnify(int scale, T element) throws MagnificationException

/**
 ...
 * @return {@inheritDoc}
 */
 <T> T magnify(int scale, T element) throws MagnificationException

Introduced as a block tag in JDK 1.0, and as an inline tag in JDK 16.

see

Adds a "See Also" heading with a link or text entry that points to a reference. A documentation comment can contain any number of see tags, which are all grouped under the same heading. The see tag has three variations. The form to reference other program elements is the most common. This tag is valid in all documentation comments. For inserting an inline link within a sentence to a package, class, or member, use a link tag.

Introduced in JDK 1.0.

serial

Used in the documentation comment for a default serializable field. See Documenting Serializable Fields and Data for a Class.

See also Oracle's Criteria for Including Classes in the Serialized Form Specification.

An optional field description should explain the meaning of the field and list the acceptable values. When needed, the description can span multiple lines. The standard doclet adds this information to the serialized form page.

If a serializable field was added to a class after the class was made serializable, then a statement should be added to its main description to identify at which version it was added.

The include and exclude arguments identify whether a class or package should be included or excluded from the serialized form page. They work as follows:

For example, the javax.swing package is marked with the @serial exclude tag in package-info.java. The public class java.security.BasicPermission is marked with the @serial exclude tag. The package-private class java.util.PropertyPermissionCollection is marked with the @serial include tag.

A serial tag at the class level overrides any serial tag at the package level.

Introduced in JDK 1.2.

serialData

Uses the data description value to document the types and order of data in the serialized form. This data includes the optional data written by the writeObject method and all data (including base classes) written by the Externalizable.writeExternal method.

The serialData tag can be used in the documentation comment for the readObject, writeObject, readExternal, writeExternal, readResolve, and writeReplace methods.

Introduced in JDK 1.2.

serialField

Documents an ObjectStreamField component of the serialPersistentFields member of a Serializable class. Use one serialField tag for each ObjectStreamField component.

Introduced in JDK 1.2.

since

Adds a "Since" heading with the specified since-text value to the generated documentation. The tag indicates that the associated program element has existed since it was introduced in the software release specified by the since-text value. For example, @since 11.

The text has no special internal structure, meaning that commas, whitespace characters and other punctuation have no special significance. Multiple since tags are allowed; for example, you could use multiple tags when the program element is part of more than one API. However, when used in conjunction with the feature in the standard doclet to generate a summary page of what API is new in each release, the since-text value of the first tag should be a simple string with no formatting markup.

It is recommended that each top-level program element (module, package, class or interface) should contain an appropriate since tag. In addition, each member element (constructor, method, field, nested class or interface) should have a tag in those cases where the value is different from that of the enclosing class or interface. If there are no since tags in the description for a member class or interface, the standard doclet will (recursively) look for any such tags in the enclosing class or interface.

Note: One or more methods may be added implicitly to a class or interface as a result of adding additional supertypes in an extends or implements clause, after the class or interface was first created.

Note: Changing the return type of a method to a covariant return type is tantamount to adding a new method with that signature, and should be given an updated since tag. This is because the method with that new return type could not be called in any prior release to leverage that new return type.

Introduced in JDK 1.1.

snippet

Includes a fragment, or "snippet", of example code in the generated documentation. The code may be provided inline within the tag by specifying a body and/or in an external file, specified in the attributes. Within the content, markup tags can be placed in line comments to identify regions within the text and instruct how to present the text in these regions.

Additional details about the snippet can be given as attributes, in the form of name=value pairs, placed after the initial tag name. An attribute name is always a simple identifier. An attribute value may be an identifier, unsigned integer, or enclosed in either single or double quote characters; no escape characters are supported. An attribute value and the preceding = may be omitted when the presence of the attribute name is sufficient. Attributes are separated from the tag name and from each other by whitespace characters, such as space and newline.

A snippet may specify an id attribute, which can be used to identify the snippet in both the API and the generated HTML, and which may be used to create a link to the snippet. In the generated HTML, the id will be placed on the outermost element that is generated to represent the snippet.

Code fragments are typically Java source code, but they may also be fragments of properties files, source code in other languages, or plain text. A snippet may specify a lang attribute, which identifies the kind of content in the snippet. For an inline snippet, the default value is java. For an external or a hybrid snippet, the default value is derived from the extension of the name of the file containing the snippet's content.

Inline snippets

An inline snippet contains the content of the snippet within the tag itself.

The content of the snippet, which is included in the generated documentation, is the text between the first newline after the colon (:) and the closing curly brace (}).

There is no need to escape characters such as <, >, and & with HTML entities, nor is there any need to escape documentation comment tags.

Surrounding whitespace is stripped from the content using String::stripIndent.

There are two limitations on the content of inline snippets:

  1. The character sequence */ cannot be used anywhere within the content, because the */ would terminate the enclosing documentation comment. This includes use of /* ... */ comments, or within a // comment, or within string literals, such as those used to represent regular expressions. This restriction applies to all content in documentation comments; it is not specific to the snippet tag.

  2. The content of an inline snippet can only contain balanced pairs of curly-brace characters. The overall inline tag is terminated by the first right brace that matches the opening brace. This restriction applies to all inline tags; it is not specific to the snippet tag.

External snippets

An external snippet refers to a separate class or file that contains the content of the snippet.

In an external snippet the colon, newline, and subsequent content can be omitted.

Unlike inline snippets, external snippets have no limitations on their content. In particular, they may contain /* ... */ comments.

The location of the external code can be specified either by class name, using the class attribute, or by a short relative file path, using the file attribute. In either case the file can be placed in a package hierarchy rooted in a snippet-files subdirectory of the directory containing the source code with the snippet tag. Alternatively, the file can be placed on an auxiliary search path, specified by the --snippet-path option to the javadoc tool. The use of snippet-files subdirectories is similar to the use of doc-files subdirectories for auxiliary documentation files.

The file for an external snippet may contain multiple regions, to be referenced in different snippet tags, appearing in different parts of the documentation.

Hybrid snippets

A hybrid snippet is both an internal snippet and an external snippet. It contains the content of the snippet within the tag itself, for the convenience of anyone reading the source code for the class being documented, and it also refers to a separate file that contains the content of the snippet.

It is an error if the result of processing a hybrid snippet as an inline snippet does not match the result of processing it as an external snippet.

Markup tags

Markup tags define regions within the content of a snippet. They also control the presentation of the content, for example highlighting parts of the text, modifying the text, or linking to elsewhere in the documentation. They can be used in internal, external, and hybrid snippets.

Markup tags begin with @name, followed by any required arguments. They are placed in // comments (or the equivalent in other languages or formats), so as not to unduly interfere with the body of the source code, and also because /* ... */ comments cannot be used in inline snippets. Such comments are referred to as markup comments.

Multiple markup tags can be placed in the same markup comment. The markup tags apply to the source line containing the comment unless the comment is terminated with a colon (:), in which case it is as if the tags were present on the immediately following line. The latter syntax may be useful if the markup comment is particularly long, or if the syntactic format of the content of a snippet does not permit comments to appear on the same line as non-comment source. Markup comments do not appear in the generated output.

Because some other systems use meta-comments similar to markup comments, comments that begin with @ followed by an unrecognized name are ignored as markup comments and will appear in the generated output. If the name is recognized, but there are subsequent errors in the markup comment, then an error is reported. The generated output in such cases is undefined, with respect to the output generated from the snippet.

Regions

A region is an optionally-named range of lines that identifies the text to be displayed by a snippet. They also define the scope of actions such as highlighting or modifying the text.

The beginning of a region is marked by either

The end of a region is marked by @end or @end region=name. If a name is given then the tag ends the region started with that name. If no name is given then the tag ends the most recently started region that does not already have a matching @end tag.

There are no constraints on the regions created by different pairs of matching @start and @end tags. Regions can even overlap, although we do not expect such usage to be common.

Highlighting

To highlight content on a line or in a range of lines, use @highlight followed by arguments that specify the scope of the text to be considered, the text within that scope to be highlighted, and the type of the highlighting.

If region or region=name is specified then the scope is that region, up to the corresponding @end tag. Otherwise, the scope is just the current line.

To highlight each instance of a literal string within the scope, specify the string with substring=string where string can be an identifier or text enclosed in single or double quotes. To highlight each instance of text matched by a regular expression within the scope, use regex=string. If neither of these attributes are specified then the entire scope is highlighted.

The type of highlighting can be specified with the type parameter. Valid type names are bold, italic, and highlighted. The name of the type is converted to a CSS class name whose properties can be defined in the system stylesheet or overridden in a user-defined stylesheet.

Modifying the displayed text

It is often convenient to write the content of a snippet as code that can be accessed and validated by external tools, but to display it in a form that does not compile. For example, it may be desirable to include import statements for illustrative purposes along with code that uses the imported types. Or, it may be desirable to display code with an ellipsis or some other marker to indicate that additional code should be inserted at that point. This can be done by replacing parts of the content of the snippet with some replacement text.

To replace some text with replacement text, use @replace followed by arguments that specify the scope of the text to be considered, the text within that scope to be replaced, and the replacement text.

If region or region=name is specified then the scope is that region, up to the corresponding @end tag. Otherwise, the scope is just the current line.

To replace each instance of a literal string within the scope, specify the string with substring=string where string can be an identifier or text enclosed in single or double quotes. To replace each instance of text matched by a regular expression within the scope, use regex=string. If neither of these attributes are specified then the entire scope is replaced.

Specify the replacement text with the replacement parameter. If a regular expression is used to specify the text to be replaced then $number or $name can be used to substitute groups found in the regular expression, as defined by String::replaceAll.

To delete text, use @replace with an empty replacement string. To insert text, use @replace to replace some no-op text placed where the replacement text should be inserted. The no-op text might be a '//' marker, or an empty statement (;).

Linking text

To link text to declarations elsewhere in the API, use @link followed by arguments that specify the scope of the text to be considered, the text within that scope to be linked, and the target of the link.

If region or region=name is specified then the scope is that region, up to the corresponding @end tag. Otherwise, the scope is just the current line.

To link each instance of a literal string within the scope, specify the string with substring=string where string can be an identifier or text enclosed in single or double quotes. To link each instance of text matched by a regular expression within the scope, use regex=string. If neither of these attributes are specified then the entire scope is linked.

Specify the target with the target parameter. The form of its value is the same as used by the standard inline link tag.

Introduced in JDK 18.

spec

Identifies an external specification in terms of its URL and title. The URL may be absolute or relative. Relative URLs will be evaluated against a "base URL".

All tags specifying the same URL must provide the same corresponding title; conversely, tags with different URLs must have different titles.

Introduced in JDK 20.

summary

Identify the summary of an API description, as an alternative to the default policy to identify and use the first sentence of the API description. The tag only has significance when used at the beginning of a main description. In all cases, the tag is rendered by simply rendering its content.

Introduced in JDK 10.

systemProperty

Identify property-name as the name of a system property. The name should be a "dotted identifier". In particular, it must not contain whitespace characters, or characters such as }. No other content is permitted in the tag; the possibility of additional content is reserved for future use. The tag can be used in all documentation comments.

Introduced in JDK 12.

throws

Adds an exception with the specified name followed by the specified description to the "Throws" section of a Method or Constructor detail.

exception-name should refer to an exception that might be thrown by the method, and should either be the name of an exception class or a type variable. This tag is valid only in the documentation comment for a method or constructor.

A documentation comment may use multiple throws tags for the same or different exceptions. If such a comment is then targeted by an inheritDoc tag, it copies all exceptions of that kind into the "Throws" section.

Note: An unchecked exception class may be omitted from the throws clause, and so won't be considered missing in a method documentation. To inherit a description of such an exception, add a corresponding throws tag with an inheritDoc tag in its description.

To ensure that all checked exceptions are documented, when a throws tag does not exist for an exception in the throws clause, the standard doclet adds that exception to the generated output (with no description) as if it were documented with the throws tag.

The exception tag is equivalent to this tag, although using a throws tag is now recommended.

Introduced in JDK 1.2.

uses

This tag may only appear in the documentation comment for a module declaration, and serves to document that a service may be used by the module. The description may be used to specify the characteristics of the service that may be required, and what the module will do if no provider for the service is available.

Introduced in JDK 9.

value

Displays the value of a static field with a compile-time constant value.

The format string may be omitted, in which case a default format will be used, appropriate to the type of the field. If the format string is given, it must either begin with a percent character (%) or be enclosed in double-quote characters ("). It must contain exactly one percent character. The string must conform to the definition of a format string, as specified in the documentation for java.util.Formatter. The conversion specified in the format string must be appropriate for the type of the constant value.

When the value tag is used without a field_reference argument in the documentation comment of a static field, it displays the value of that constant:

/**
 * The value of this constant is {@value}.
 */
public static final String SCRIPT_START = "<script>"

When used with a field-reference argument in any documentation comment, the value tag displays the value of the specified constant:

/**
 * Evaluates the script starting with {@value #SCRIPT_START}.
 */
public String evalScript(String script) {}

Introduced in JDK 1.4; format added in JDK 20.

version

Adds a "Version" subheading with the specified version-text value to the generated documents when the -version option is used. This tag is intended to hold the current release number of the software that this code is part of, as opposed to the since tag, which holds the release number where this code was introduced. The version-text value has no special internal structure.

A documentation comment can contain multiple version tags. When it makes sense, you can specify one release number per version tag or multiple release numbers per tag. In the former case, the standard doclet inserts a comma (,) and a space between the names. In the latter case, the entire text is copied to the generated document without being parsed. Therefore, you can use multiple names per line when you want a localized separator other than a comma.

If there are no version tags in the description for a member class or interface, the standard doclet will (recursively) look for any such tags in the enclosing class or interface.

Introduced in JDK 1.0.

Where Tags Can Be Used

The following table summarizes which tags can be used in which contexts.

Contexts in which tags may be used
Tag Module Package Type Constructor Method Field Other
author ✓ ✓ ✓ ✓
code ✓ ✓ ✓ ✓ ✓ ✓ ✓
deprecated ✓ ✓ ✓ ✓ ✓
docRoot ✓ ✓ ✓ ✓ ✓ ✓ ✓
exception ✓ ✓
hidden ✓ ✓ ✓
index ✓ ✓ ✓ ✓ ✓ ✓ ✓
inheritDoc ✓
link ✓ ✓ ✓ ✓ ✓ ✓ ✓
linkplain ✓ ✓ ✓ ✓ ✓ ✓ ✓
literal ✓ ✓ ✓ ✓ ✓ ✓ ✓
param ✓ ✓ ✓
provides ✓
return ✓
see ✓ ✓ ✓ ✓ ✓ ✓ ✓
serial ✓ ✓ ✓
serialData *
serialField *
since ✓ ✓ ✓ ✓ ✓ ✓ ✓
snippet ✓ ✓ ✓ ✓ ✓ ✓ ✓
summary ✓ ✓ ✓ ✓ ✓ ✓ ✓
systemProperty ✓ ✓ ✓ ✓ ✓ ✓ ✓
throws ✓ ✓
uses ✓
value ✓ ✓ ✓ ✓ ✓ ✓ ✓
version ✓ ✓ ✓ ✓

Notes: