Sometimes, a spontaneous road trip can be a lot of fun, as long as you’re willing to take the good with the bad—getting lost, car trouble, unfriendly (or just plain weird) natives, bad diner food. Usually, though, the most successful trips involve planning, roadmaps, and best of all, guidance from people who’ve already been there.
The journey from traditional, deliverable-centric content creation to DITA-based content creation falls into this second category. In this session, we talk about one small publication group’s experience moving to DITA, from the initial discussions to the successful implementation of a FrameMaker-based, end-to-end publication process. Here are some of the high points of the project; we’ll discuss our decision-making process and some of our technical approaches in detail in the session.
Raspberry Pi 5: Challenges and Solutions in Bringing up an OpenGL/Vulkan Driv...
The Trip to DITA
1.
2. About the Trip to DITA
The transition from traditional, document-centered content creation to DITA-
based content creation requires substantial planning for the journey to be
successful.
The journey of one small documentation team took from unstructured Frame to
Structured FrameMaker using the DITA content model. This will include
initial discussions, strategic development and steps required for a successful
implementation of a FrameMaker-based, end-to-end publication process.
3. About this Presentation
An overview of the migration from unstructured to structured authoring
A specific process for preparing FrameMaker documents for conversion to
DITA.
Challenges faced during the year-long journey from unstructured to
structured authoring.
A model other documentation teams can use to re-organize existing
documents and to create new documents going forward to help ensure their
FrameMaker content is well-positioned for a successful, efficient
conversion effort.
4. Interviews and Report
Interviews with team members and group
discussions
Focus on business drivers, team readiness, content
analysis, tools and disparate processes.
Offered an opportunity to discover team members’
preparedness.
Report outlined summary of findings with
requirements and action plan for a phased
implementation.
Recommendations, detailed development steps
and basis of agreement.
5. Goals
Optimize authoring, translation and publishing
cycle to help meet increased need for translations
Reuse content from one publication to another
quickly and consistently
Decrease time to develop and update manuals
Automate labor-intensive tasks giving writers
more time for research and improving content
Prepare for a future Content Management System
(CMS) implementation
6. Budget
Client workgroup gathered information for a year
prior to contacting consultant. Spread expenses over
two years and four phases.
Phase 1: Assessment of business goals,
requirements analysis, content audit, content
analysis, process mapping and process re-
engineering
Phase 2: DITA training, content modeling,
prototype project, assess content management
technologies
Phase 3: Write a CMS RFP based on business
requirements, distribute RFP to potentials vendors,
proof of concept with qualified vendor(s)
Phase 4: Buy and implement CMS
7. Content Audit
Writers selected three publications representative
of their work products
Consultants worked with client team to performed
content audit
Determine how well content could be broken
into topics
Determine what rewrite required
Develop method for determining reuse
Spreadsheet for tracking
Selected book with most reuse (the beta book,
80%) and used that for development
8. Content Modeling
DITA was selected as the content model
Content modeling focused on chunking topic by
types and identifying attributes audience, product,
language, software version and compliance
requirements
Used content audit spreadsheet as a guide
Content modeling was a client training exercise
9. Tool Selection
Hard copy and PDF formats only
Primary requirement is PDF for print
Writers had no experience with structured
authoring
Writers already experienced Frame users
Migrated from unstructured (7.1) to structured
9
Completely Frame based publishing process
preferable to an XSL-FO publishing process
Using default DITA functionality, not DITA-FMx
Yeah, we know FMx is better!
10. Reuse
Based on content audit:
Text: 85%
Graphics: 70%
Three products, four document types
Import XML software file as conrefs for variables
Conditions
11. Clean Template
Standardize style names, eliminate un-used styles
Clean and standardize all templates before
conversion
Client made good decisions to simplify template
for translation:
No hyphenation
No text in cross-references
12. Conditions
Eliminated conditional text and replaced with
@product, @audience (pick list built in EDD)
Structure changes needed to accommodate
Decisions made about level of granularity for
these attributes
topic-level
element-level
sentence level?
phrase level?
More tracking required with last two levels
13. EDD
Based on “default” DITA EDD that ships with
FrameMaker
Unnecessary rules removed or simplified
Client uses only ~40% of tagset but unused
elements not deleted from EDD
Use of @outputclass on several elements to
specify formatting not predictable via context
Full-page images vs in-column images
Ad hoc table cell alignment
No specialization or DITA plug-ins
14. Conversion Table
Template standardization
Template style usage very consistent—YAY!
“Crosswalk” spreadsheet developed to map
existing template styles to equivalent DITA
elements
15. Document Conversion
Conversion table “wrapped” every instance of a
format in the appropriate element
Higher-level wrapping largely manual—writer
discretion needed to determine appropriate topic
type, etc.
Converted each chapter document into one large
XML doc
Consultants/writers then determined how to “split”
large doc into individual topics
Naming convention
Not a mass conversion
Consistent level of granularity
16. Variables
Original template contained 100+ variables
corresponding to menu options in software
Value of variables came from developer-generated
XML file
Easily translatable; kept this functionality
Created separate file containing each variable as a
conref source (<varname>)
This file still developer-generated, though now
in DITA
Still easily translatable
17. Images
Most images in original documents were line
drawings of screen overlaid with text boxes
containing variables, simulating screen shots
Easily translatable, reduced need for language-
specific images
Initial goals was to use SVGs (Scalable Vector
Graphics) to maintain this functionality (i.e.
translatable text as element within SVG)
FrameMaker cannot “look inside” SVG code; this
solution not workable at present
For same reason, cannot show/hide callouts
(Frame cannot interpret the @display value)
Use SVGs to lay groundwork for CMS
18. Images
Image sizing
Frame uses DPI, converts to @height, @width
No units in XML code
@height, @width not accessible in Frame
Image paths
Frame uses backslashes by default
Cannot enter backslashes manually, must
change to forward slashes
Re-linking image via Missing File dialog does not
dynamically update @href path
19. Translation
Prefix for <note> elements, some <title> elements
and other items language-specific
Decision made to generate these items using EDD
rules based on @xml:lang
One template per document type vs. separate
template for each language
Non-English PDFs produced by the translation
agency, so this simplification was essential
Numbering/bulleting not language-specific but
groundwork laid if changes in future
20. Tables
Table handling
Mapping between DITA table elements/
attributes handled by read/write rules
Trick is learning what corresponds to what in
Table Designer and Custom Ruling & Shading
dialogs
Most formatting applied based on <tgroup>
not <table>
<colspec>, <spanspec> in XML code but not in
Frame—can be confusing
21. Revisit Template
Client focused on page layout
Ad hoc tables modifications required narrowing
Impact on EDD
Revisions along the way; discovery of
“exceptions”
Writers rethinking items in light of increasing
DITA knowledge
22. Beta Project
Representative of typical book but not a
production project
Used to test every aspect of development from
topic types, templates, EDD, conversion, reuse,
book building and training.
Smaller sample of beta book for translation testing
Project book is leveraged for production
23. AXCM
Free Frame plug-in from West Street Consulting
www.weststreetconsulting.com
Filters files for @product, @audience using XPath
filters
Colors documents to emulate conditional text
Filters topics on the fly
More robust than Frame’s built-in Filter By
Attribute
For this publishing process, preferable to applying
a ditaval file
24. Basic Publishing Process
Ditamap for each book chapter
Save ditamap as composite Frame file
Set numbering, other doc properties
Not using “Prompt for Ditaval” option
Not using bookmap
Add composite Frame files to Frame book
Filter files for @product, @audience using
AXCM
Generate/update TOC, Index
Publish book as PDF
25. Installation
Upgrade to FrameMaker 9
Set up structured application
Install AXCM
26. Demonstration
Three days on-site to rollout template
Goals:
Instruct writers in DITA authoring
Get writers familiar with tagset
Get writers familiar with Frame’s DITA
implementation and interface
Explain conversion process
Additional day on-site to instruct writers in basics
of EDD creation and maintenance
Writers not ready to take over EDD but they
understand what’s “under the hood”
27. Workflow Requirements
CMS is just one part of a unified content strategy
Specification detailing client use case
Requirements focus
Support for DITA, content reuse, user
interface, support encoding for languages,
translation streaming and branching capacities,
automated, multi-channel publishing
Cost
Criteria for migration
Use to produce a Request for Proposal (RFP)
include information about existing
implementations, licensing, configuration, core
technologies, maintenance, support,
customization, training and cost
28. Request for Proposal
RFP is offered to major CMS vendors and vendors
are given the opportunity to respond.
Evaluate responses
Based on requirements, select vendors who
respond to the RFP with a proof-of-concept
29. Proof of Concept
Beta book used for Proof of Concept
Vendors use client content to demonstrate how
their system meets client needs
Compare Proof of Concept results with CMS
Requirements
Evaluate vendor relationship, expertise
30. Contact Information
Mollye Barrett
ClearPath, LLC
www.clearpath.cc
mollye@clearpath.cc
www.linkedin.com/in/mollyebarrett
Leigh White
ElementalSource, LLC
http://elementalsource.pbworks.com/
elementalsource@gmail.com
www.linkedin.com/in/leighwwhite