Schema based transformation: architectural forms for the 21st Century

John Cowan

Should be able to transform from the document we have to the document we want.

Source document. Source schema with decorations. Target schema. Form or renaming attribute. Names of renaming attributes to be ignored.

Element renaming. Add attributes with target element names in the source elements.


Mixed content.

Have to look at this paper. Could be useful.

Decision making in XSL-FO formatting

Tony Graham

No decision control once FO is produced. Fire and forget processing. They have formatting features requiring layout decisions. Preform at tables in three widths on long pages, compare area trees, use width with least area and no overflow. There’s no standard XML format for area tree. SVG? For graphics, image magical identify for size and resolution.

XSL-FO charter has run out. There’s a FOP extension for working with AreaTree. Use Saxon-CE event model for XSL-FO? I encouraged him to create an extension for working with AntennaHouse AreaTrees.

Fat Markup: Trimming the myth one calorie at a time

David Lee

Is XML really fat? Tests, 2 JSON variants, 3 XML variants. XML better file size, compressed size about the same. Parsing speed similar. Query speed similar. Use HTTP compression, usually. JS libraries sacrifice performance for ease of use. Choice of markup matters, but not for the performance reasons that people usually say.

General Architecture for Generation of Slide Presentations

Eliot Kimberly

PowerPoint from DITA. Slide systems, PowerPoint, Apple, Slidey and other HTML systems. Transform in convenient way. Transform from arbitrary XML to intermediate to slide presentation format. XSLT from source to SlideSetML, java from SlideSetML to non-XML formats, XSLT to XML formats. Uses POI to generate PPTX.

Last night out


Unexpectedly received an invitation to visit a great Montreal game shop with a bunch from the conference, so I had to go. Afterwards, Wendell Piez and Simon St. Laurent headed back while the rest of us had dinner at La Prunelle (pictured here Dave Dubin, Jacob Jett, Matt Patterson, and Syd Bauman). The conversation was great throughout and so was the food (in my case, escargot in a pastry shell, veal rib in a white wine sauce with zucchini and chanterelles, and a fine creme brûlée). Stayed up way too late — I think it’s got to be coffee for me in the morning, if I’m going to make it through the sessions at all. Still, a perfect night.

Transcending triples in semantic modeling

Micah Dubinko

(SPARQL keeps getting mentioned. Need to look that up. )

1. Limit the universe down to a particular set of triples. 2. Query the triples.

What does it mean for a triple to be embedded in a document?

Wait. Was there a point? I must have missed something. Or a lot.

Good, one of the questioners said he thought it seemed unfinished too.

The new W3C publishing activity

Liam Quin

Open web platform. Publishing, commercial and academic, not tweeting, blogging, etc. The web isn’t designed for professional publishing. W3C publishing activity. Should I try to get in on the paged media activity?

Interactive XSLT in the browser

O’Neil Delpratt and Michael Kay

Not until 2008-2009 could you reliably use XSLT 1.0 in the browser. Saxon-CE. Showing interactive web app using XSL. Never noticed the event stuff in XSLT 2.0 before. People are jazzed. They apparently can do SVG also.

I will probably be trying this.

First Person: WebVTT versus TTML

Andreas Tai

Subtitles like half a tweet. Formatting, timing. Used to get transmitted during the vertical blank interval. Europe uses ebu-stl binary format. Timing, alignment. Limited formatting, file size, character sets. Two alternatives, TTML and WebVTT. TTML XML based, supported by broadcasters and video players. WebVTT text based, supported by browser vendors. QC important to broadcasters, so XML. Web more about error handling. (Seems like this could easily have been avoided.)