Microsoft's OOXML spec was so much larger than ODF for at least 3 reasons.
1. OOXML is an XML-ization of a descendent of earlier binary office formats. You can do things in binary formats that don't map efficiently to XML. That resulted in OOXML having uglier and more verbose XML. If I recall correctly this was only a small part of the size difference.
2. OOXML was way more complete. This was a big part of the size difference.
The first approved standard for ODF for example gave no details on how formulas worked in spreadsheets. All you could get from the standard was that you should support formulas.
The OOXML spec on the other hand described in detail the format of formulas and gave detailed specs for a large number of functions. There were often 2 or 3 pages for a single function.
3. When IBM and Sun were touting that 6000 page number for the draft OOXML spec, they were using a printing that has something like half as many lines per page as the printing of ODF they were comparing with.
The other side to that is that MS appeared to "stuff" ISO committees in various places to force it through, and didn't provide enough detail in their huge spec for true interop.
Perhaps neither was perfect, but it certainly seemed like MS played dirty.
1. OOXML is an XML-ization of a descendent of earlier binary office formats. You can do things in binary formats that don't map efficiently to XML. That resulted in OOXML having uglier and more verbose XML. If I recall correctly this was only a small part of the size difference.
2. OOXML was way more complete. This was a big part of the size difference.
The first approved standard for ODF for example gave no details on how formulas worked in spreadsheets. All you could get from the standard was that you should support formulas.
The OOXML spec on the other hand described in detail the format of formulas and gave detailed specs for a large number of functions. There were often 2 or 3 pages for a single function.
3. When IBM and Sun were touting that 6000 page number for the draft OOXML spec, they were using a printing that has something like half as many lines per page as the printing of ODF they were comparing with.