Field Guide
Why S1000D Does Not Come With a PDF
The Short Answer
A Data Standard, Not a Page Design
S1000D standardizes content and its exchange: data modules in a common source database, assembled into publications by publication modules. The official description says that data storage system enables output in both page-oriented and interactive electronic formats. It does not say the specification hands you the output.
Look at what the Issue 6 and Issue 7 download packages contain: the specification, schemas, sample data, a default BREX, a front matter sample set, a functionality matrix, and a decision point index. A stylesheet that turns data modules into a PDF is not on the list. The specification does have guidance for page-oriented publications, and a good stylesheet starts from it. But guidance is not a stylesheet, and the choices it leaves open are settled by each project.
What the Data Holds
Structure Yes, Pages No
- Structure and order
- A publication module lists data modules, other publication modules, and external publications in order, under titled entries that nest. That gives a formatter the table of contents and the reading order. It does not give it a page.
- Front matter as data
- The front matter schema has a title page, a table of contents, and lists. A table of contents entry can carry a number of pages, but that number only exists after something has paginated the publication.
- Change marking as data
- Elements can be flagged as added, modified, or deleted, with a change mark switch and references to the reason for update. The data says what changed. A change bar in the margin is drawn by the stylesheet.
- A few print hints
- Tables can be marked portrait or landscape and page-wide. Graphics can carry a reproduction width, height, and scale. These are hints a stylesheet may honor, not a layout.
Everything else about a page (size, margins, fonts, running heads, where a break falls, what number a page carries) is produced by the pipeline. That is the same bargain described in how XML becomes a PDF: content without appearance, and a stylesheet that manufactures the appearance.
The Pipeline
From Publication Module to PDF
- Read the publication module and pull in the data modules it references, in order.
- Filter for the configuration being printed, if the project filters.
- Transform the assembled content into a page description. The classic route is XSLT producing XSL-FO; a CSS-based paged media route does the same job with a different language.
- Run a formatter to set lines, tables, and pages, and write the PDF.
- Post-process: resolve page references, build lists of effective pages, apply change bars, and check the result.
Steps 3 to 5 are where a project's print rules live. Viewers face the same assembly and filtering work, with different output; see what an S1000D viewer has to do.
Why Every Project Differs
Print Rules Belong to the Business Rules
S1000D expects each project to record its decisions in business rules. The Issue 4.2 grouping, which our business rules examples walk through, includes a data output category, and the BREX carries the machine-checkable part. Print decisions sit there: page size, which front matter is required, how warnings are placed, how change is shown.
Customers add their own. A program may need its new PDFs to resemble a legacy manual, or to meet a service-specific style. Legacy formatting often lives in the old pipeline's style files; see what a FOSI is for the SGML-era version of the same problem. The result is predictable: two programs on the same S1000D issue can need very different stylesheets.
Where It Breaks
Six Problems Every Print Pipeline Meets
- Applicability filtering
- One data module can serve several configurations, with applicability on individual elements. A print run has to decide: filter to one configuration before formatting, or print everything with applicability statements. Both are legitimate; the choice changes page counts and cross-references, so it has to be settled before the stylesheet is written. Applicability, with an example.
- Warnings and cautions
- Warnings and cautions are structured elements, and steps can point at them by reference instead of repeating them. On paper that means the right warning must land before the step it governs, on the same page, every time, even when the step moves. Keeping a warning with its step is a pagination rule, and each project writes its own. Warnings, cautions, and notes.
- Illustrated parts tables
- A parts catalog is a figure plus long, indented rows. Rows that do not fit a page, wide tables that need landscape, and the figure that should sit near its items are where formatters struggle most. Expect this part of the stylesheet to take the longest. Illustrated parts data.
- Graphics
- A data module points at an illustration by an entity reference. If the reference does not resolve, the picture silently disappears from the PDF. The official sample set ships its illustrations in CGM, so a print pipeline often needs a conversion step before the formatter ever sees them. Information control numbers.
- Many modules, one book
- A publication is dozens or hundreds of short modules. Where does each start: a new page, a new chapter, or run on? How are pages numbered, and how does a reference to another module turn into a page number? Page references need the pagination to be finished first, so they are usually resolved on a second pass. The publication module.
- Effective pages and change bars
- A list of effective pages, with each page's issue and date, is a product of the formatted output, not of the data. So are change bars. Both depend on comparing this issue with the last one at the page level, which means the pipeline has to keep the previous issue's result. Plan this on day one, not at the first revision. The CSDB.
Buying and Contracting
What to Ask For
- Name the S1000D issue the stylesheet must support, and say whether a move to a newer issue is in scope.
- State the page size, the output formats (print-ready PDF, screen PDF, or both), and who owns the layout rules, in writing.
- Require the stylesheet and all supporting files to be delivered with the publication, with the right to modify and reuse them.
- Specify which front matter is generated, which comes from data modules, and how the list of effective pages is produced.
- Say how change is shown on a revision: change bars, a highlights list, or both.
- Decide the applicability rule for print: one configuration per PDF, or all configurations with statements.
- Require the project's BREX and business rules to cover the data output rules, and to be agreed before production starts.
- Define an acceptance sample: a set of real modules, including a long parts table and a warning-heavy procedure, rendered and signed off before the full run.
- State how rendering errors are reported, and who fixes a data problem versus a stylesheet problem.
Read the text of your own contract and business rules for what is already required; this list is a prompt, not legal advice. For the wider toolchain decision, see choosing S1000D tools and free S1000D tools.
FAQ
Questions We Hear
Can S1000D produce a PDF?
Yes, but not by itself. S1000D defines the data and how it is exchanged. A PDF comes from a pipeline that assembles the publication, filters it, applies a stylesheet, and runs a formatter. The official download does not include that stylesheet; you build it, buy it, or adapt a community one.
Why isn't there one standard S1000D stylesheet?
Because the specification leaves presentation to the project. Customers have different page sizes, house styles, and legacy looks to match, and the business rules record those decisions. A stylesheet written for one program's rules rarely fits the next without changes.
Does S1000D use XSL-FO?
The specification does not require it. XSLT feeding XSL-FO to a formatter is the long-established route to page-oriented output, and CSS-based paged media is the modern alternative. Either can work. What matters is that the stylesheet is delivered, documented, and covers your hard cases.
Are there free S1000D stylesheets?
A few open-source and community options exist, and some editors offer PDF output as an add-on. Check which S1000D issues each one supports, when it was last updated, and its license, because coverage often stops at an older issue. Our guide to free S1000D tools lists what we could verify.
Why does my S1000D PDF drop images or tables?
The usual causes are an unresolved illustration reference, an illustration format the formatter cannot read, or a table that does not fit the page and has no break rule. Test with a few of your own difficult modules before judging a pipeline on a clean sample.
Get In Touch
S1000D Output That Prints Right?
Bring a few of your hardest data modules and your business rules. Start from what the print output has to do before anyone writes a stylesheet.