[mpeg-OTspec] Re: Apple Spliced Font Format

[James Cloos]

>>>>> "TT" == Tony Tseung <tseung at apple.com> writes:

TT> There was a talk of the name types being a number rather than the
TT> name, this was simply in the merits of the name table specification
TT> of TrueType (OpenType) fonts, the enum value are identical to the
TT> id's in the name table. I meant to make a list of aliases but didn't
TT> quite get arround for that. It may be good to spell out the name type.

I recognized that the id numbers were from the SFNT name tables.  And I
do not see any problem in the regard with your current xml as an internally
defined, proprietary format.

But given how verbose XML was designed to be, when it (XML) is used in
an international standard, it should be as self-documenting as possible.

The (preferably RelaxNG) schema should provide all of the documentation
needed to correctly implement the standard, whether one is creating the
xml files or executing them.

My comments were only about how the concept should be specified by the
forthcoming standard, not criticism of your current, internal implementation.

A similar issue is the question of putting values in attributes or as
children of the element.  Eg, which of these may be preferable:

         <Name type="FullName" name="Symbol" language="en"/>
vs:
         <Name type="FullName" language="en">Symbol</Name>

Or, similarly, which of:

	<language language="en"/>
vs:
	<language>en</language>

Such a difference has been common discussion point in a number of XML-based
standardizaions.  Each model has its own idiosyncrasies; it would be useful
to discuss the cross-platform effects of each when developing the standard.

-JimC
-- 
James Cloos <cloos at jhcloos.com>         OpenPGP: 1024D/ED7DAEA6