ISO/IEC JTC 1/SC34 N0xxx


Information Technology --
Document Description and Processing Languages
-- Information Presentation

TITLE: Liaison Comment to JTC1/SC29/WG11
SOURCE: Toshiya Suzuki/Hiroshima University
STATUS: Informal Liaison
ACTION: For information
DATE: 2011-03-30
DISTRIBUTION: SC34, SC34/WG2 and Liaisons

Liaison Comment to SC29/WG11 about Panose value in ISO/IEC 14496-22


Importance of the abstract typeface information

In various editable or final form data like ISO/IEC 29500 (OOXML), PDF, PCL, or the supplementary data like CSS, the referential informations of the font are written and important to recover the fonts when the font itself is not embedded in the document. In most case, the name of the font face is used as the unique identifier, but sometimes the document rendering system does not have the font with the same name. In such case, the abstract specification of the typeface, with the informations of typographic characteristics of the font is important to substitute the missing font by the appropriate font that the rendering system can use.

Discussion in SC34

In OOXML, following properties are written in the documents.
Data indirectly generated from the fonts
  • Pitch property used in Microsoft Windows GDI
  • Family property used in Microsoft Windows GDI
  • charset property used in Microsoft Windows GDI
Data directly generated from the fonts
  • supported Unicode range
  • Panose
The data indirectly generated from the fonts are controlled by the document producer, the request of the validation is easy. The document including invalid pitch, family and charset are recognized as invalid OOXML document. The supported Unicode range is a collection of the boolean values for each range of the scripts in Unicode, so there is no invalid data. The last one, Panose-1 is usually described as 10 hexadigits or 10 digits. In the case of TrueType and OpenType, it is stored as 10 octets. In comparison with the space of possible values of 10 octets (00 00 00 00 00 00 00 00 00 00 - FF FF FF FF FF FF FF FF FF FF), the defined valid range of Panose is quite smaller. Thuhere was a long discussion in SC34/WG4 how to validate the Panose values and what kind of invalid values are popular. During the discussion, it is found that there are some documents including insufficient or misleading information about the Panose definition and it is supposed that many font developers have not checked genuine Panose spec. The genuine Panose spec have per-family-kind properties (e.g. Latin Symbol family kind does not have a property for serif style), but some insufficient or misleading documents describe all properties are independently defined and the properties for Latin Text are generic (thus they described as Latin Symbol family kind have a property for serif style). It seems that the earliest one would be MSDN document, and the insufficient description was duplicated to the earlier OpenType spec. The latest OpenType and ISO/IEC 14496-22 spec are improved, it does not have the text about the valid value range of each properties. But the name of properties are still included. They are same with the properties of Latin Text, and misleading the developers to the conventional misunderstanding that the propertis are independent with the family kind. SC34/WG2 proposes to drop the list of the property names from next version of ISO/IEC 14496-22 and just refer genuine Panose spec.

Expected impact of proposed change

Recently, Microsoft participants to SC34/WG4 stated that the existing implementation in Microsoft products follow to genuine Panose spec, but MSDN document insufficiently described. Therefore, it is expected that there is no change in Microsoft products.

Other comments

Although SC34/WG2 proposed to reduce the description about Panose in future version of ISO/IEC 14496-22, SC34/WG2 strongly propose to keep the text "International: Additional specifications are required for PANOSE to classify non-Latin character sets."