Optimizing Content Reuse with DITA - LavaCon Webinar with Keith Schengili-Roberts

Optimizing Content Reuse with
DITA
Keith Schengili-Roberts
DITA Specialist, IXIASOFT

Agenda
• Introduction
• Content Reuse in DITA: The Mechanics
 A look at the tools built into DITA for creating reusable
content
• If You Can’t Find it, You Can’t Reuse It
 How to make your content more findable so that it can
be reused
• Content Reuse in DITA: How to Make it Work
 Making reusable content efficient and findable
• Q/A

Who’s This Guy?
Keith Schengili-Roberts,
IXIASOFT DITA Specialist
What I do:
• DITA evangelist
• Liaison with OASIS; on DITA
Adoption, DITA Technical and
Lightweight DITA Committees
• Industry researcher
• Lecturer on Information Architecture,
University of Toronto
• 10+ Years of DITA XML experience

Also Known As “DITAWriter”
• Industry blog started +5
years ago
• Just over 215,000 hits
• Regularly updated info on:
 DITA Conferences
 DITA Books
 Companies Using DITA
 DITA CMSes
 DITA Editors
 Other DITA Tools
 DITA Consulting Firms
• News/views on DITA use
• Features interviews with
those making a difference
in the world of DITA

CONTENT REUSE IN DITA:
THE MECHANICS
A look at the tools in DITA for creating reusable content

Reuse is Built-in to DITA
• DITA was built around
the idea of content
reuse
 This has helped make
DITA the fastest
growing XML-based
technical
communications
standard (over 630
companies world-wide
have adopted DITA)

Content Reuse = Consistent Messaging
• Benefit of consistent content and messaging
• Consistent content means consistent user
experience
 Along with being seamless, available and context-
specific

Reduced Localization Costs with DITA
• Content reuse in English = localization savings
• If there are many target languages, ROI
argument for the move to DITA (+ CCMS) is
easier

DITA-related Localization Savings Example

The Levels of DITA Reuse
• Structural (ditamap, map)
• Topic
• Conref (and Conkeyref, and Conref “push”)
• Conditional processing
• Keys
• DITA 1.3: branch filtering and keyscopes
Other:
• Special considerations for reusing images
• Metadata to help find content to reuse
 more on both of these later

Matryoshka (Russian Doll) Reuse
• With each of these
levels, it is possible to
add the next sub-level
of reuse
• So a reused map might
(for example) contain a
topic from another map,
which in turn uses a
conref, conditional
processing and keys

Structural Reuse
• Topmost level of reuse, involving reuse of entire
ditamaps and maps
• ditamaps: goes hand-in-hand with either
conditional processing and/or keys
• maps: ditto above, but when used as “chapters”
can also be published as separate documents
 IXIASOFT’s documents follow both scenarios

DITA Metrics Tangent: Ratio of
Structural Elements
• Can search for maps and bookmaps
being used and can determine a ratio
from it
• Why? Provides an idea as to how
content is being structured
• If, for example, you use maps as “sub-
maps”, would expect to see more maps
than bookmaps
 That’s exactly what we see here
• IXIASOFT uses “sub-maps”
extensively, as we can work on smaller
aspects of our maps, to better organize
larger maps, and to reuse sub-maps.
Our DITA CMS User Guide includes a
dozen sub-maps.
• For small documents, such as release
notes, bookmap chapter structure is
overkill
Count: 118

Topic Reuse
• At its easiest, it is simply
taking a topic and using it
elsewhere
• Can be done at the
topicref level, or by using
a conref (+ key)
• May also involve use of
conditions and keys used
within the map

Conrefs
• Involves taking paragraph(s), sentences or phrases
 Handy when you just want to reuse something smaller
than a whole topic
• Requires addition of an ID to the conref being targeted
 This should be coordinated with others
 Should use a standard style/naming convention

Avoid Spaghetti Conrefs!
• Avoid having nested
conrefs (i.e. conrefs that
contain conrefs), or
conrefs that could be
subject to change
• IXIASOFT DITA CMS
includes a specialization
designed specifically for
holding conref-able
content
 Can do the same thing by
simply agreeing to have
conref-able content on
designated topics

Conref-ing Down the Rabbit Hole
• There’s reuse, and then
there’s ridiculous levels
of reuse: ex. you could
conref every single letter
in your topics
 There comes a point where
making the conref takes
more time than simply
writing “new” content
• In general, don’t conref
below the level of a
phrase

Conrefs and Localization
• Conrefs can pose a risk to translation
because reusable content is
translated without the context of the
topics in which it is used
• Translators cannot process sentences
that are segmented by phrase-level
conrefs
 They typically require a PDF with all
conrefs resolved so that they have the
context necessary for translating the
content
• Best practice: conref only the
following
 Keywords (such as product names)
 Block-level elements and NOT parts of
blocks (like phrases)
• Given all this, why not just use keys
instead?
Problematic Example:
<p>If you revert a referable-content topic that’s
reached Localization:done, then all the topics
that reference that topic will have their status
set to their initial Localization state. <ph
conref=" topic123.xml#topic123/abc">These
topics will be included in the localization kit.
</ph>Their content will remain unchanged and
they need not be sent to the translation
team.</p>
Best Practice Example:
<p conref=“topic123.xml#topic123/abc“>If you
revert a referable-content topic that’s reached
Localization:done, then all the topics that
reference that topic will have their status set to
their initial Localization state. These topics will
be included in the localization kit. Their content
will remain unchanged and they need not be
sent to the translation team.</p>

Conkeyrefs (Conref “Pull”)
• These are indirect content
references defined by
keys at the map level
• Good for pulling in short,
term values referenced
elsewhere
 It “pulls” in values referenced
elsewhere
• Pro: Change terms easily
by pointing to different
values in the referred topic
• Con: referred topic needs
to be managed over time
• Here are some conkeyref targets in a
table:
• Which can then be referenced within
topics:

Using conkeyref to Hold a Product Term
Key Vebulon 5000 benefits
<term id="ph_prodname">Vebulon 5000
</term>
Map with keydef topic highlighted
Reference to the keydef term in another topic:
<title>Key <term conkeyref="r_productname_variables/ph_prodname"/>
benefits</title>
How this is expressed at output:
• Easy reuse emerges when you use change product name to
something else (e.g. “Vebulon 6000”)

Best Practice: Hold Your Conrefs / Keys in a Table
• Put your content references (terms, phrases, and images) in
a topic designed to hold them, and use a two column table
that both describes the reference and what it resolves to
• All product terms in one place, all images in another, etc.

Conref Push
• Designed to “push” content
at a marked place
• Pro: perfect for a “quick fix”
where you need to insert
something extra, like an
extra step to handle
regional variances for
example
• Con: arguably trickier/less
straightforward than keys;
without good
communication can lead to
unwanted surprises for
other writers
• Content from task topic:
1. Remove the old oil filter.
2. Drain the old oil.
3. Install a new oil filter and gasket.
4. Add new oil to the engine.
5. Check the air filter and replace or clean it.
6. Top up the windshield washer fluid.
But for cars sold in North Carolina (for
example), we need to include an extra step:
3. Recycle the motor oil. In Durham, North Carolina,
you can take it to the Waste Disposal and Recycling
Center.
Conref push allows you to add this
intermediate step within the original task
topic

Conditional Processing (Profiling)
• Allows conditions to be
applied within topics so
that elements can be
included or excluded
during output
processing: DITAVAL is
your friend!
• Pro: great for creating
content aimed at
multiple audiences
• Con: can become
unwieldy as more
conditions are added
Sample DITAVAL file:
<val>
<prop action="exclude"/>
<prop action="include" att="audience"
val="everybody"/>
<prop action="include" att="audience"
val="novice"/>
<prop action="include" att="product"
val="productA"/>
<prop action="include" att="product"
val="productB"/>
</val>
This example excludes all
values not specified in the
DITAVAL, and includes
audience=everybody and
novice plus
product=product and
productB

Recommended Best Practices for
Conditional Processing
• Write your conditions as you create you content
 And keep track of what you use!
 A good internal Style Guide will state commonly used
conditions and expected values
• Don’t go overboard: if things appear to be
getting too complex, they probably are.
Consider writing separate topics
• Test your output! There’s nothing worse than
revealing content not intended for a particular
audience / product

Using ditavalref for Branch Filtering
• In DITA 1.3 conditions can
be set within the map and
then filtered using a
ditavalref (which links to a
ditaval file with specific
conditions set)
• Allows for greater flexibility
when publishing content; if
for example you have
install instructions for
Windows, MacOS and
Linux editions of your
software, you can use
Branch Filtering to render
them at one go
• Previously, with just ditaval you
could only publish one install
set per OS:
Win MacOS Linux
• With branch
filtering this content
can be combined in
a single pass
Win
MacOS
Linux

Keys
• An indirect
addressing
mechanism where
key names are
attached to
resources which are
then referenced
elsewhere
• Very flexible, highly
adaptable
file.dita:
<topic id="topicid">
<title>Example referenced
topic</title>
<body>
<section id="section-
01">Some content.</section>
</body>
</topic>
Key definition defined in the map:
<map>
<topicref keys="myexample"
href="file.dita"/>
</map>

Best Practice for Keys
• Eliot Kimber’s opinion: “use keys everywhere”
• Basically there are two types of keys: resource
and navigation
 Resource keys link to URLs, images, etc., contained in
“warehouse” topics
 Navigation keys (or “normal keys”), which are named
key for topic links
 Eliot also recommends using keys with conrefs (conkeyref)

DITA Reuse Metrics
• “Reuse of DITA Topics? What is the Best Metric to
Measure the Success of Your Reuse of DITA
Topics?” (http://ow.ly/X7mzM) by Bill Hackos

Bill Hackos’ Reuse Formula
• Percent Repository Words Reused in Context =
(Words in All Produced Content – Words in the
Repository) / (Words in the Repository)
%
Words =

Example Based on IXIASOFT
DITA Documentation for 2015
Based on 2015 numbers:
• Total number of words in the repository:
268,663
• Words in All Produced Content: 623,078
• PRWRC = (623,078 – 268,663)/268,663
• PRWRC = 354,415 / 268,663 = 132%

How is a 100%+ Value Possible?
• Easy: ditaval
• Though ditaval is not mentioned in original article, Bill
Hackos does talk about +100% values being entirely
possible
• We have number of publications that are created based on
a series of ditaval value, as much as 21 per bookmap
• Percentage value is
the lower bound on
our content reuse;
could be considerably
more but is no less
than 132%

IF YOU CAN’T FIND IT,
YOU CAN’T REUSE IT
How to make your content more findable so that it can be reused

Need for Effective Reuse Grows with Time
• Results of a recent
survey where people
were asked how many
topics are in a given
doc project
• While peak is between
100-500 topics, many
are considerably larger
• But you need to find
content in order to
reuse it!

Finding Content for Reuse: Taxonomy
• Check to see if your
company or industry
already uses a
taxonomy
• If you don’t already have
one, brainstorm a
taxonomy with your
colleagues
• Example “mind map”
was used for revamping
dita.xml.org website,
using MindMeister.com

Inserting Metadata into Topic and Maps
• Resulting MindMap can be
outputted to text, then
converted to XML
• Individual elements can then
be used within a Subject
Scheme map, (or whatever
your CMS supports) so you
can select elements to insert
as metadata values in maps
and topics
• When added to maps,
metadata is inherited at
topic level
In DITA source:
Expressed in XHTML at output:

Advantage of CMS Object Approach
• When it comes to
finding your own
content, some CMSes
(such as the IXIASOFT
DITA CMS) associate
additional metadata
with content in the
associated property
object for each DITA
file in the system

Finding Content for Reuse
• Solution: devise effective metadata and tag your
topics accordingly!
• Use descriptive titles
• Use short descriptions!

CONTENT REUSE IN DITA:
HOW TO MAKE IT WORK
Making reusable content efficient and findable

The Human Angle
• So far we have only talked about the mechanics,
but it is people who have to make this work (with
or without benefit of a CCMS)
• Past the mechanics, effective reuse within a
documentation team comes down to:
 Writing content for reuse
 Finding content to reuse
 Communicating about reusable content

Moving from Legacy Content
• Few who move to
DITA have the luxury
of starting fresh;
legacy content
needs to be
converted
• Do an audit of this
material, look for
what is common
• Port only the content
that will be used in
the future

Create a Reuse Strategy
Once the content audit is
done, think about:
• Naming conventions for
filenames, IDs, etc.
• Determine how granular
reuse will be: topic,
element or phrase
• What criteria to use
when content should be
forked
Add this material to your
DITA Style Guide!

Communicate About Reusable Content
• Create metadata to
describe your content to
make it findable
• Determine where and
what reusable content
will be stored
• Coordinate about how
and when content can
be reused within the
team
Prevent this from happening!

Don’t Go Overboard
Evaluate reuse efforts
against the returns:
• It is just easier to
simply write “Click OK”
rather than conref-ing it
“Apples to apples, dogs
to ducks”
• When things are very
different, don’t try to
squeeze reuse out of
them

Knock Down Those Silos!
• Effective reuse means
that content can no
longer be siloed
• By necessity, content
“ownership” needs to be
shared and
communicated
 One person or group
should not “own” their
content, and change to
content that is already
shared needs to be
coordinated

Writing Content for Reuse
• Keep things concise:
content reuse and
minimalism go hand-in-
hand!
• Omit terms like “see the
following” or “the above
image…” etc.
• Think about reuse
possibilities as you write
 DITA content is inherently
modular, so drop any
pretense of narrative

Further Reading
• “DITA 1.2 feature article: conref push”, by Kris Eberlein
(dita.xml.org/resource/dita-12-feature-article-conref-push)
• Thunderbird DITA demo files: https://github.com/gnostyx/dita-
demo-content-collection
• “The Elusive Promise of Content Reuse” by Leigh White (article:
ow.ly/XS0qs, SlideShare: ow.ly/XS0vb)
• “Managing Deliverable-Specific Link Anchors: New Suggested
Best Practice for Keys (YouTube: ow.ly/XS0gz)
• Taxonomy Warehouse: www.taxonomywarehouse.com
• “@conref @conkeyref @conrefpush: Reuse Strategies in DITA
When Migrating Legacy Content” by Adam Sanyo ow.ly/XS0db
• “Reuse of DITA Topics? What is the Best Metric to Measure the
Success of Your Reuse of DITA Topics?” by Bill Hackos
ow.ly/X7mzM
• oXygen DITA usage survey: http://ow.ly/4nkLpa (prelim. results:
http://ow.ly/4nkLuh)

QA
• Don’t forget to attend my presentation at LavaCon
Las Vegas!
 Improve Your Chances for Documentation Success
with DITA and a CCMS, Wed October 26 @ 9:30am-
10:30am
• Blog: www.ixiasoft.com/en/news-and-events/blog
• Twitter: @IXIASOFT (and @KeithIXIASOFT)
• IXIASOFT DITA CMS Users LinkedIn group:
www.linkedin.com/groups?gid=3820030
• Member of OASIS DITA Technical Committee

Conref-ing Down the Rabbit Hole Example
• In this code example,
the pangram “The
quick brown fox jumps
overs the lazy dog”
uses a conref for each
and every letter in the
sentence
• Just because you can
do something doesn’t
mean you should ;)

A conref/conkeyref Example Using a Table
• Which can then be
referenced within
individual topics:
• The conkeyref points to
the filename/id
(productname_variables.di
ta + id="ph_prodname"
used in the table
• Here are some conkeyref
targets in a table:

Using conkeyref for Holding a Product Name
= Key Vebulon 5000 benefits
From r_productname_variables.dita
Map with keydef topic highlighted
keydef filename
(minus “.dita”)
term id being
referenced
In a separate
topic:

Creating an Image Warehouse Using Keydefs
keydef for warning
icon image set in a
separate file
(images-key.ditamap)
Arranging the images in a
table
(r_image_warehouse.dita)
Sample row from table linking warning icon to an ID
Image is referenced/displayed in a topic using conkeyref

Conref Push
• Designed to “push” content
at a marked place
• Pro: perfect for a “quick fix”
where you need to insert
something extra, like an
extra step to handle
regional variances for
example
• Con: arguably trickier/less
straightforward than keys;
without good
communication can lead to
unwanted surprises for
other writers
• Example from Kris Eberlein’s article on conref
push at: http://dita.xml.org/resource/dita-12-
feature-article-conref-push
Original topic with id added:
Conref “push” topic that points to id in original
topic, then specifies content to include:
Resulting output:

Using ditavalref for Branch Filtering
• In DITA 1.3 conditions can be set within the map and then
filtered using a ditavalref (which links to a ditaval file)
• Example below shows install section where ditavalref processes
each OS branch (win/mac/linux)
• In this case, the contents of the
installing branch appear three times,
once for each referenced ditaval
(where each platform name is set to
include exclusively), so one set of
installation instructions per platform
(win, mac, and linux)
• For mac content processed as
XHTML, install.dita and do-stuff.dita
will be processed with mac values
included, and filename will be have “-
apple” appended to them; because
mac-specific.dita is processed in the
mac pass, its contents are included
here, but excluded in the win and
linux processing passes

Optimizing Content Reuse with DITA - LavaCon Webinar with Keith Schengili-Roberts

More Related Content

What's hot (20)

Viewers also liked (6)

Similar to Optimizing Content Reuse with DITA - LavaCon Webinar with Keith Schengili-Roberts (20)

More from IXIASOFT (20)

Recently uploaded (20)

Optimizing Content Reuse with DITA - LavaCon Webinar with Keith Schengili-Roberts

Editor's Notes