This patent details an idea whereby instead of storing documents as one large file that combine both content (e.g., text) and formatting intermixed, they instead separate the content from the formatting markup into separate files, allowing content to be changed independently of style. The exact structure of the separation is not detailed in the patent, just the concept of separation.

IBM systems have been using this concept for decades, so IBM's prior art could be used to invalidate the patents of i4i and Microsoft. I first encountered it during our 1998 Y2K conversion from ERP on IBM S/36 to IBM AS/400, which was first released some time before that, but I believe the concept was used earlier on IBM S/38.

Check out IBM's "external definitions" of files tables, and other objects, where the definition says what layout or separation between fields or columns of a table, and what can go in each of them, like dates, text, floating point, integers, images, urls, or whatever. This separates the content, from the formatting, to simplify software development.

Previously, on IBM systems, every program needed a separate set of definitions of file content (internal definitions), and different programs could have different incompatible definitions.

What about seminal DTP programmes? I used Xerox Ventura Publisher back in 1989 running on a PC/AT. You created your text in a Word processor/text editor, and laid it out/formatted it in Ventura. Your stylesheet defined the formatting applied to the "tags" you used which, like XML, were custom and extensible. I used simple macros in Word (for DOS) to pre-format a lot of my text - all you had to do was prepend the tag name in front of the paragraph. Very XML-like. In fact the Ventura UK user group's newsletter was even called "Tagline".

My memory is a little hazy on this, (it's a long time ago) but I think that there was a .chp file which did most of the layout work, and a .Pub file (which just defined the order of chapters in a publication. formatting and tag names). You could have any number of text files (Word, Wordperfect etc) and illustrations/images.

It was a very elegant architecture and I was heartbroken when Corel acquired Ventura and messed it up by creating a monster chapter that, instead of simply referencing all the constituent files, embedded them, a la Word for Windows. Tragic!