Features
Four components, from authoring constraints to multi-format publishing, forming one continuous pipeline.
DITA / S1000D structured authoring engine
Built on DITA 1.3 and S1000D, with topics or data modules as the smallest manageable content units. The XML schema constrains structure at the moment of writing: <cmd> accepts imperative sentences only, <warning> mandates risk, consequence, and mitigation, and <shortdesc> generates a module abstract automatically. In S1000D environments, the BREX module enforces terminology, warning format, and sentence-length compliance; DMRL verifies data module completeness; SNS assigns a unique identifier to every data module for full lifecycle traceability.
22 Chinese authoring rules · real-time validation
Twenty-two Chinese authoring rules are embedded in Oxygen XML Author across five dimensions: terminology accuracy, format compliance, technical rigor, user-friendliness, and safety compliance. Key rules include: R1 mandatory terminology-mapping-table lookup; R15 one action per sentence; R14 one topic per paragraph with a seven-line limit; R19 separation of action and result in procedural steps; R21 safety warnings requiring risk scenario, consequence, and mitigation; R16 one term per concept throughout. Violations surface at the point of writing, not in a post-hoc batch check.
ASD Simplified Technical English (STE) rule set + AI plugin
English output quality is embedded at the Chinese authoring stage through STE rules. The AI translation engine ingests structured content, terminology bases, and STE rules, enforcing term matching, translation-memory reuse, and sentence-length and style constraints during translation. Chinese source text is written for STE translatability; translation quality is built in before translation begins.
Integrated architecture
Single DITA/S1000D source → structured authoring with real-time validation → AI translation (CN→EN→RU, etc.) → multi-format publishing to PDF, HTML5, WebHelp, and IETM. The same content asset stays consistent across output formats without duplicate maintenance.