Design Science Proposal to Microsoft to Help
STEM (Scientific/Technical/Engineering/Mathematical) Publishers
Work with Office 2007 Documents

August 28, 2007

Paul Topping
Design Science, Inc.
www.dessci.com

Top scientific publishers like Science and Nature currently do not accept author submissions in Microsoft Word 2007's new document format. Their publishing workflows are designed to handle earlier Word formats containing MathType or Equation Editor equations, rather than the OMML (Office Math Markup Language) equations. Other changes in Word's document format also cause difficulties. This proposal describes several features to be added to the next version of Design Science's MathType for Windows product that will help these publishers better handle author submissions in Office 2007's new document format. These features will also help authors who use Office 2007 collaborate with other authors that use an earlier version of Office. Microsoft's assistance is requested in support of this effort.

The main feature to be added to MathType is conversion of OMML equations to MathType's equation format. This requirement is driven by the following, in order of importance:

1. Scientific publishers' workflows rely on MathType's ability to save equations as EPS, GIF, or MathML.
2. Some publishers have editorial departments that rely on MathType to edit equations.
3. Authors that prefer to use MathType as their equation editor may want to collaborate with others that prefer OMML.

The conversion of OMML equations to MathType format will be implemented within the MathType product itself. However, there are a number of issues that require Microsoft's assistance and, possibly, bug fixes and small enhancements to Word 2007. Mostly, the need for assistance stems from the fact that OMML equations make extensive use of Microsoft's new Cambria Math font. For this reason, MathType needs to be able to support Cambria Math or some characters in OMML equations may not convert properly. Cambria Math has a few characteristics that make it a unique font: it makes extensive use of Unicode Plane 1 codepoints and contains special math formatting features. These are supported by a new, undocumented OpenType MATH table.

MathType equations are currently saved as graphical images in Windows Metafile Format (WMF). Unfortunately, WMF is a very old format and doesn't support Unicode at all, let alone Plane 1, and, therefore, most characters in Cambria Math cannot be displayed in MathType's WMF equations. Microsoft Windows superseded the WMF format many years ago with the Enhanced Metafile Format (EMF), which does support Unicode. Unfortunately, MathType cannot produce its equations in the EMF format until Microsoft Word implements a mechanism for aligning such equations with the text baseline in Word documents. Currently, this baseline information is communicated to Word and other applications via a comment in the WMF data. It could be done this way in EMF as well or, even better, as a new COM interface for the purpose which would be supported by Word. We request that Microsoft implement a new baseline alignment mechanism in Word that will allow MathType to produce equations as EMF.

Presumably, it is Microsoft's intention to create other math fonts that make use of the new math formatting mechanisms introduced with Cambria Math. In addition, as OpenType is a public standard, other font vendors should be allowed to create math fonts that implement the extended standard and applications, such as MathType, should be allowed to support them. In fact, the STIX organization has expressed interest in making their fonts compatible with OMML equations. This requires, as far as we know, that their fonts make use of the new OpenType MATH table technology. In addition, a recent presentation by Microsoft's Murray Sargent at the TypeCon typography conference implies that a new Windows component, MathFont.dll, may also be required to support Cambria Math and other fonts built to this new standard. We request that Microsoft make public all information on their new MATH font tables and other OpenType extensions implemented in Cambria Math. We also request that any software required to work with the new fonts be freely redistributable by application vendors. Finally, we request that Microsoft make any licensing terms required to use this information and software as liberal as possible.

While MathType 6.0 generally works well with Office 2007, our customers have reported a number of small problems that affect their work. Details can be provided. We request that Microsoft fix these problems:

• In certain circumstances, dragging a MathType equation from one table cell to another crashes Word.
• Characters typed immediately to the right of a previously created MathType equation are inserted with an incorrect baseline.
• MathType commands moved from the MathType Tab on Word 2007's Ribbon to the Quick Access Toolbar sometimes disappear later.
• When MathML is present on the clipboard, Word's Paste Special should offer "MathML Presentation" as an option and choosing that option should create an OMML equation.

Design Science expects to continue to support STEM publishers and their authors, as well as scientists, engineers, and educators in general, with software tools for mathematical publishing. We hope that Microsoft will continue to support us and our mutual customers in that effort. A positive answer to the requests outlined in this paper will go a long way to helping in that effort. We look forward to Microsoft's response.