Converting StarOffice & StarWriter (SDW) to Markdown: OLE Compound Binary Forensics

Before LibreOffice and Apache OpenOffice became household names, Star Division (and later Sun Microsystems) produced StarOffice, one of the earliest cross-platform office suites for Windows, Linux, and Solaris. Its word processor, StarWriter, saved documents with the .sdw extension.

Decades later, millions of government, legal, and academic documents from the late 1990s remain locked in .sdw files. Modern office suites have dropped support for older StarWriter filters. In this article, we examine how .sdw compound files work and how convrtr extracts clean GitHub Flavored Markdown (GFM) directly in your browser.

---

1. The OLE 2.0 Compound File Binary (CFB) Architecture

StarWriter 5.x documents are built on Microsoft's OLE 2.0 Compound File Binary format (magic bytes 0xD0 0xCF 0x11 0xE0 0xA1 0xB1 0x1A 0xE1). The file acts like a miniature filesystem inside a single file:

1. Header: 512 bytes containing sector size parameters, FAT sector locations, and directory pointers.
2. FAT Sectors: Sector allocation chains mapping multi-sector streams.
3. Directory Stream: 128-byte entries defining root, storages, and streams.
4. Data Streams:
   • StarWriterDocument: The primary binary document payload containing formatted paragraphs, tables, and styles.
   • \x05SummaryInformation: Standard OLE property set holding document title, author name, subject, page count, and word count.

---

2. Text Run Extraction and Control Code Sanitization

Inside StarWriterDocument, textual paragraphs are stored in sequential binary chunks delimited by control words, formatting tokens, and font markers.

To convert this data to clean Markdown:

1. FAT Traversal: Reads the sector chain starting at the directory pointer to assemble the complete stream in memory.
2. Metadata Recovery: Reads \x05SummaryInformation to extract properties like title (Property 2) and author (Property 4) for YAML frontmatter.
3. Run Filtering: Traverses byte buffers to extract printable ASCII and Latin-1 text runs while discarding font table definitions and binary formatting codes.
4. Structure Recognition: Detects headings based on line length and case, parses bullet characters (*, -, 1.), and structures the output into valid Markdown.

---

3. How to Convert SDW to Markdown Online

1. Drag and drop your .sdw file into SDW to Markdown.
2. convrtr parses the compound directory, decodes property streams, and reconstructs the document body into Markdown in milliseconds.
3. Download your clean .md file or copy the Markdown directly into Obsidian, Notion, or GitHub.