Introduction
Why is MetadocXML Needed?
MetadocXML is not needed, it's an additional resource for the Gentoo
Documentation Project to keep track of documents, even if they are located
outside of the normal [gentoo]/xml/htdocs/doc scope.
Thanks to MetadocXML, we can now
-
track documents that are located inside project webspaces
(/proj) instead of the usual documentation repository
(/doc)
-
categorize documentation into various categories (or subcategories) with the
additional benefit that we can now automatically generate the documentation
index (and more)
-
track unofficial documentation team members (such as translators)
-
use parts of big documents (Handbooks) as individual guides on certain
topics
-
assign bugs to particular documents for quick reference and with the
possibility of masking out a document in case of a major showstopping bug
-
primitively check if a translated file is in sync with its English
counterpart or not
Note that the last advantage is primitive and will probably not be extended.
Some translation teams use scripts based on trads-doc to manage
translations, others use online translation management tools. If you are
starting up translations for Gentoo, pop by on #gentoo-doc or ask the
gentoo-doc@gentoo.org mailinglist for help.
Translation teams that do not use MetadocXML yet don't need to worry - they will
not lose any current functionality as it only builds upon the existing
infrastructure - there are no changes to the GuideXML format that need
MetadocXML.
How does MetadocXML Work?
There is one central file in which all meta information on the documentation is
maintained. We call this file metadoc.xml. This file should be
located inside your main repository (/doc/${LANGUAGE}) although
this is not hard-coded.
Inside this file, all meta information is stored:
- Members of the team
- Categories in which documents participate
- Files that are covered
- Documents that are covered
- Bugs that are part of a document
Next to metadoc.xml, one also can have a dynamically generated index
file (usually called index.xml), an overview listing of all
documentation (usually called list.xml) and an overview listing of
all members, files and bugs (usually called overview.xml).
The metadoc.xml File
XML Structure
The metadoc.xml file is started with the usual XML initialisation
code and Gentoo CVS header information:
<?xml version='1.0' encoding="UTF-8"?>
<!-- $Header: /var/cvsroot/gentoo/xml/htdocs/doc/en/metadoc.xml,v 1.25 2004/12/23 09:51:30 swift Exp $ -->
<!DOCTYPE metadoc SYSTEM "/dtd/metadoc.dtd">
Then, one starts with the MetadocXML declaration.
<metadoc lang="en">
Translators should reference the main /doc/en/metadoc.xml in the
parent attribute. This lets metadoc identify untranslated files and find
out whether versions of translated versions and originals still match.
<metadoc lang="language code" parent="/doc/en/metadoc.xml">
Beneath the metadoc entity, the following entities should be declared (in
the given order):
-
version to help keep track of changes
-
members which declares all members of the given language team
-
categories which declares the possible categories used
-
files which contains all files covered by the Metadoc file
-
docs which contains all documents covered by the Metadoc file
The Version Entity
The version number should be increased when a document or a file is added or
removed, when a path is changed or on any update that might have an impact on
translated versions.
The Members Entity
Inside the members entity, one can declare two 'types' of members: lead
and member. A lead should be known by the Gentoo Developers
Relations as it takes only the nickname of the Lead developer and looks it up in
the Gentoo Memberlist. A member can either be a Gentoo Developer (in
which case only a nickname is given) or a contributor.
In case of a contributor, a member tag is given two attributes,
mail and fullname, containing the contributor's e-mail address and
full name.
<members>
<lead>swift</lead>
<lead>neysx</lead>
<member>dertobi123</member>
<member mail="gentoo_contributor@gmail.com" fullname="John Doe">jdoe</member>
</members>
The Categories Entity
Inside the categories entity one only declares cat entities. Each
cat entity covers one Category. It uses one mandatory parameter id
which is used to reference the category. You can also define a parameter
parent in case the category is a child of another category.
In this case, the parent attribute references the id attribute of
the parent category.
<categories>
<cat id="faq">Frequently Asked Questions</cat>
<cat id="install">Installation Related Resources</cat>
<cat id="install_guides">Installation Guides</cat>
</categories>
The Files Entity
The files entity contains only file entities.
Each file entity references a single XML file. It has a mandatory
id attribute which should be seen as a primary key to lookup the file.
Metadoc will compare the file name defined with the same id attribute in
the metadoc's parent file (defined in the root element) to find out whether the
file is a translation or an untranslated file. File names would be identical in
the latter case.
The metadoc file itself can be listed and will appear on the overview page.
<files>
<file id="metadoc">/doc/en/metadoc.xml</file>
<file id="ati-faq">/doc/en/ati-faq.xml</file>
</files>
The Docs Entity
The docs entity should only contain doc entities.
Each doc entity has a mandatory fileid attribute, which refers to
the id attribute of a file entity corresponding with the main file
for the document.
In case of a handbook chapter, the doc entity must contain a
bookref entity which references the main handbook page (the top handbook
XML file). This entity then contains two attributes, called vpart and
vchap which refer to the corresponding part and chapter of the document
inside the handbook.
Inside the doc entity, two other entities are possible:
-
One or more memberof entities, referring to the category or
categories in which the document is located (note that a document can be in
several categories at once)
-
One bugs entity containing one or more bug entities. A
bug entity refers to a bugnumber that covers a bug in the document.
In case of a major bug, one can add the attribute stopper="yes" to
the bug entity in order for the document not to appear on the
generated index page.
sysadmin_specific
102481
1151330
sysadmin_general
]]>
Example metadoc.xml file
The Gentoo site uses a metadoc.xml file to aggregate information of
all its documentation. You can view the current version one online.
The Additional MetadocXML Files
Automatically Generated Index
When you want an automatically generated index, you should start the document
with the following code:
<?xml version='1.0' encoding='UTF-8'?>
<?xml-stylesheet href="/xsl/metadoc.xsl" type="text/xsl"?>
<?xml-stylesheet href="/xsl/guide.xsl" type="text/xsl"?>
<!-- $Header$ -->
<!DOCTYPE dynamic SYSTEM "/dtd/metadoc.dtd">
<!-- Substitute "/doc/en/metadoc.xml" with the location of your metadoc file -->
<dynamic metadoc="/doc/en/metadoc.xml">
<title>Gentoo Documentation Resources</title>
<intro>
(...)
</intro>
<catid>faq</catid>
<catid>install</catid>
</dynamic>
In between the intro tags you should write one or more sections which
will always appear on the top of the page. You will probably want to write an
introduction and some additional information for the reader to know who to
contact in case of translation mishaps or other issues.
Inside the intro tags you can use plain GuideXML starting from
section.
The catid tags refer to the main categories used by the dynamical index.
You should list each possible non-child category that is declared in your
metadoc file. Do not list categories that are children of another
category.
Dynamically Generated List Document
A dynamically generated list document starts identically to a dynamically
generated index file:
<?xml version='1.0' encoding='UTF-8'?>
<?xml-stylesheet href="/xsl/metadoc.xsl" type="text/xsl"?>
<?xml-stylesheet href="/xsl/guide.xsl" type="text/xsl"?>
<!-- $Header$ -->
<!DOCTYPE dynamic SYSTEM "/dtd/metadoc.dtd">
<!-- Substitute "/doc/en/metadoc.xml" with the location of your metadoc file -->
<dynamic metadoc="/doc/en/metadoc.xml">
<title>Gentoo Documentation Listing</title>
However, there is no intro tag. What needs to be added are all the top
categories used by the listing. To differentiate this from the index (which will
also display the abstract information on each document) this happens between
list tags inside listing:
<listing>
<list>faq</list>
<list>install</list>
</listing>
Dynamically Generated Overview Document
The overview document is started similarly as the two documents decribed above:
<?xml version='1.0' encoding='UTF-8'?>
<?xml-stylesheet href="/xsl/metadoc.xsl" type="text/xsl"?>
<?xml-stylesheet href="/xsl/guide.xsl" type="text/xsl"?>
<!-- $Header$ -->
<!DOCTYPE dynamic SYSTEM "/dtd/metadoc.dtd">
<!-- Substitute "/doc/en/metadoc.xml" with the location of your metadoc file -->
<dynamic metadoc="/doc/en/metadoc.xml">
<title>Documentation Development Overview</title>
You can again write up a small introduction in GuideXML between the intro
XML tags, starting from a section up. Once that is finished, a single tag
<overview/> is sufficient.
<intro>
(...)
</intro>
<overview/>