One consistent arguement on why OOXML should not be approved as a ISO standard is poor design. Read Rob Weir’s blog entry on what XML design is about, and that will give you a good understanding of what the goals, engineering decisions and trade off. I should know: I designed not one but a few XML format.
Weir’s last point on consistency of XML representation, using text color and text alignment as an example, is something I have deeply feeling about. In one of my first XML design conducted a good few years ago, I was conscious on consistency, but invariably introduce some inconsistency in the design. Those inconsistency hurts me till today.
However, I disagree with him on one point. When he says “The question we should be asking is not whether a standard is similar to an application’s internal representation. That doesn’t really matter.”, he was wrong. While this is technically true from an engineering point of view, if an XML representation reflects too closely an application’s internal representation, it makes adoption by others more difficult. XML design should be application-neutral, otherwise you just put barriers between your XML design and adoption by others. Call it professional jeolousy by your detractor or whatever you want. It exists.
Even with me, where I always try to divorce personal interest when evaulating something, if I see some XML that too closely resembles an application’s internal representation, I have alarm bells ringing in my head. For a start, that design is flawed from an engineering viewpoint. At the bare minimum, it shouts “the designer had not managed to separate application-specific perculiarity from the design”. That normally escalate quickly to “the designer had designed the XML for one application and does not give any consideration to general adoption”. At times, after reading the whole design, one have to conclude that the design is too entrenched in one application and it is not worthwhile rework it for general adoption.