I am currently working on the import of shapes from OOXML into Writer. As the title of this post suggests, you might think that importing shapes should be straight forward.

And using ODF it is. The following XML particle encodes a rectangle, e.g.:

<draw:rect text:anchor-type="paragraph" draw:z-index="0" draw:style-name="gr1"
    draw:text-style-name="P1" svg:width="1.8543in" svg:height="1.1567in"
    svg:x="1.2764in" svg:y="0.9744in">
    <text:p/>
</draw:rect>

This encoding is common for all of OpenOffice.org's applications. Thus, there is only one piece of code responsible for importing shapes in OpenOffice.org.

When we started importing shapes from OOXML the Impress team already was able to import some shapes from OOXML files produced by PowerPoint 2007. Consequently, we thought we could just reuse their importer code and do some adjustments and that would be it. As one might guess, life is different: If you insert a rectangle shape in PowerPoint 2007 you will get some XML like this:

<p:sp>
    ...
    <p:spPr>
        <a:xfrm>
            <a:off x="2000232" y="1500174"/>
            <a:ext cx="3429024" cy="2000264"/>
        </a:xfrm>
        <a:prstGeom prst="rect"><a:avLst/></a:prstGeom>
    </p:spPr>
    <p:style>
        <a:lnRef idx="2">
            <a:schemeClr val="accent1"><a:shade val="50000"/></a:schemeClr>
        </a:lnRef>
        <a:fillRef idx="1"><a:schemeClr val="accent1"/></a:fillRef>
        <a:effectRef idx="0"><a:schemeClr val="accent1"/></a:effectRef>
        <a:fontRef idx="minor"><a:schemeClr val="lt1"/></a:fontRef>
    </p:style>
    <p:txBody>...</p:txBody>
</p:sp>

This is DrawingML as described in chapter 5 of the Markup Language Reference for OOXML.

Copy the shape and paste it into a word document and you get this in the according DOCX:

<a:graphic xmlns:a="http://schemas.openxmlformats.org/drawingml/2006/main">
    <a:graphicData uri="http://schemas.openxmlformats.org/drawingml/2006/lockedCanvas">
        <lc:lockedCanvas xmlns:lc="http://schemas.openxmlformats.org/drawingml/2006/lockedCanvas">
        ...
            <a:sp>
                ...
                <a:spPr>
                    <a:xfrm>
                        <a:off x="2000232" y="1500174"/>
                        <a:ext cx="3429024" cy="2000264"/>
                    </a:xfrm>
                    <a:prstGeom prst="rect"><a:avLst/></a:prstGeom>
                </a:spPr>
                <a:txSp>...</a:txSp>
                <a:style>
                    <a:lnRef idx="2">
                        <a:schemeClr val="accent1"><a:shade val="50000"/> </a:schemeClr>
                    </a:lnRef>
                    <a:fillRef idx="1"><a:schemeClr val="accent1"/></a:fillRef>
                    <a:effectRef idx="0"><a:schemeClr val="accent1"/></a:effectRef>
                    <a:fontRef idx="minor"><a:schemeClr val="lt1"/></a:fontRef>
                </a:style>
            </a:sp>
        </lc:lockedCanvas>
    </a:graphicData>
</a:graphic>

Looks pretty similar to the PowerPoint XML. Our “One peace of code for one kind of thing” approach seems to hold. But, if you use Word to insert a rectangle into a Word document (DOCX), you end up with this:

<w:pict>
    <v:rect
        id="_x0000_s1026"
        style="position:absolute;margin-left:83.65pt;margin-top:16.45pt;width:249.75pt;
            height:107.25pt;z-index:251659264"
        fillcolor="#4f81bd [3204]"
        strokecolor="#f2f2f2 [3041]"
        strokeweight="3pt">
        ...
    </v:rect>
</w:pict>

This is VML as described in chapter 6 of the Markup Language Reference for OOXML. The Markup Language Reference tags VML as a deprecated format in OOXML, which is only included to the standard for backward compatibility reasons. Despite, Word 2007 uses VML to store shapes.

So what do we do? As VML is to be considered deprecated in OOXML, one might say: “Do not care about it. Use DrawingML.” If Word 2007 was only a beta release and a final version would abandon VML, that would be the approach to follow. But, the XML above is produced by the current product version of Word 2007. Customers do require that one can use Word and Writer interchangeably. It looks like we have to implement both: VML and DrawingML.

The example above is only one that depicts a more general problem. The designers of ODF had a file format in mind, that describes data. Hence, when the format describes data with the same semantics, it uses the same syntax. OOXML seems to be designed with the application model in mind. There may be different syntaxes for the same semantics, if it fits the already present application model better. But, if you want to create an alternative implementation for the format, this introduces additional effort.