DII ODF workshop catch-up

Wow - it seems the entire cast of Microsoft bloggers are laying the grounds for a whole bunch of blog-entries after the DII-workshop in Redmond the day before yesterday (Wednesday July 30^th).

So OK - I'll bite.

I am again back in Denmark after a Hell-ish evening after the workshop where jet-lag almost had the end of me. If the cap-driver had actually known the way to my hotel (he didn't so I ended up giving directions in a town I had only been in for about 36 hours in a country I was only visiting for the second time) I am sure I would have been catching Z's big time on my way home in the cap ... at 20.15 in the evening.

All in all it was very interesting to attend the workshop in Redmond. The day started up with an introduction by Peter Amstein about the approaches and decisions Microsoft had done when working on their ODF-implementation. One of the more interesting discussions here was what Microsoft had done where there was differences between ODF and OOXML - should they be conservative or creative? An example of this was numbering/bulleted lists (it is one of the key parts with differences between ODF and OOXML).OOXML has the possibility of having each bullet in a list a seperate colour where ODF does not. Microsoft had chosen the conservative approach and simply removed all colours from the bullets, but Patrick Durusau noted that he thought it was possible to use styling to do it in ODF - only downside was, that to his knowledge, this particular way of doing it was not supported by any ODF-implementation. I guess some times you are screwed regardless of what you do.

After this followed various project managers that demonstrated how their part of Microsoft Office supported ODF and talked about the remaining work. They each had complex documents that they showed us their work on and the result of saving them (OOXML-files) to ODF. I remember working with conversion using the SourceForge translator in Fall 2007, and I was really impressed by the fidelity of the conversions we saw.

Key points from this part was

• Shapes are converted without much loss of fidelity
• Metadata is converted (Dublin Core)
• Fields are converted
• Headers/footers are converted
• TOC etc are converted
• Shapes are converted
• SmartArt is converted to shapes
• Images are converted including cropping etc
• New 3D-shapes are reduced to the closest possible OpenDocument Drawing shape
• Old 3D-shapes are converted
  
• Formulas in spreadsheets are implemented by the ECMA-376 specwith Microsoft's namespace
  
• Mathematical content is converted from/to OMML and MathML
• When loading an ODF-spreadsheet
• Tables in OOXML-presentations are converted to shapes thereby making a "virtual table" since ODF does not support tables in presentations
• Conversion of embedded objects is not fully supported
• CustomXML is converted to "flat Xml" with content controls being discarded.
• When loading a spreadsheet from ODF with formula-namespace other than Microsoft's, just the values are being converted and the formulas are disgarded
• Animations are converted in presentations
• The concept of "master pages" is not converted to ODF

The rest of the day consisted of hands-on labs (stay tuned for the tech-stuff from this part of the day) and a round-table discussion in the afternoon. I will talk about these part in a couple of posts in the beginning of next week.

I saw that Doug Mahugh will post the material presented to us - so watch his blog for these data.