HTML WG Last Call comments on "CSS3 module: Ruby"

Dear CSS Working Group (and www-style subscribers),

The following comments are the Last Call comments from the HTML
Working Group on "CSS3 module: Ruby", at:

    http://www.w3.org/TR/2002/WD-css3-ruby-20021024/

I will send purely editorial comments separately later.

==========

2.2 What is ruby?

- After Figure 2.1.2, the spec says as follows:

    In the first example, a single annotation is used to annotate
    the base sequence. This simple case is typically referred as
    a 'group' ruby.

  The term "group ruby" here looks a bit confusing.  The Ruby Annotation
  spec used to call "complex ruby" as "group ruby", and changed that term
  according to comments from Japanese typography experts (during Ruby
  Annotation Last Call).  Now the term "group ruby" is used for "simple
  ruby".  Although the Glossary in Ruby Annotation REC includes the term
  "group ruby", Personally I've never heard that term in Japanese typography -
  it's sometimes called "taigo ruby" (per-word ruby), though.

  In fact in the Ruby Annotation spec, the term "group ruby" only appears
  in the glossary section and nowhere else.  Maybe it should have been
  removed from Ruby Annotation.

  Anyway, unless it's really necessary, it would be better to avoid
  the term "group ruby" here.  The usage of "group ruby" in fugures
  3.2.2 and 3.2.3 looks confusing.

- In the following paragraph, the term "mono-ruby" is used without
  explanation.  People familiar with Japanese typography can make
  sense of it, but it would be helpful to add it to the glossary,
  like Ruby Annotation REC.


3.1 Ruby specific 'display' property values

- In the first paragraph, "HTML markup" should read "XHTML markup"
  (or "XML markup").  Ruby Annotation REC explicitly removed
  "HTML markup" for ruby found in earlier drafts.

- In the last paragraph, it says:

    Conforming CSS3 user agents may not implement these ruby-related
    display' property values if they only support XHTML applications.

  We wonder if it's appropriate to pre-define CSS conformance for
  "XHTML applications" like this, without reservation.  There may
  be various kinds of "XHTML applications", and some of them might
  want to require support for CSS3 Ruby module as part of its
  conformance (say, XHTML-based simple markup language for typographic
  data exchange, similar to JIS X 4052).  We'd prefer to remove this
  paragraph.


3.2 Ruby box model

- In the paragraph after Figure 3.2.3, it says

    In the example above, the rtc element preceding the rbc element ...

  but at least in the case of XHTML ruby markup, the rtc *element* will
  never precede the rbc element, though, arbitrary XML markup could
  be defined so that an rtc-equivalent element can precede an
  rbc-equivalent element.  Anyway, to explain ruby box model, it would
  be clearer to say "ruby text" rather than "the rtc element" and
  "ruby bases" rather than "the rbc element", or, make sure to explain
  that the word "preceding" refers to the visual order and not
  necessarily the logical order of elements.

- In the second bullet of the list of exceptions, it says:

    * the equivalent of the cells: the rb element and the rt text element
      can only contain inline-level elements. This also means that a ruby
      element cannot be embedded into another ruby element.

  We don't quite understand why "[t]his also means".  In the first bullet
  of this list it clearly says that the ruby box is an inline element,
  so the second sentence is not the corollary of the first sentence.
  Ruby Annotation explicitly imposed that restriction (i.e. prohibition
  of nesting of ruby), and if CSS3 Ruby imposes similar restriction,
  fine, but that should not be just because rb and rt can only contain
  inline-level elements.

- In the last paragraph, it says:

    Finally, a conforming UA must not display the content of the rp
    element [RUBY]. The purpose of that element is to allow pre-existing
    UAs to parenthesize ruby text content.

  We think this is outside the scope of this specification.  This spec
  doesn't define any specific display value for "rp" (or anything
  equivalent), and its rendering should be determined according to its
  display value.  An XHTML 1.1 user agent may define something like

    rp { display: none }   
  
  in its user agent default style sheet, but this spec would not be
  an appropriate place to mandate that.  Otherwise, what's an "rp"
  element?  Is it any element in any namespace with the local part "rp"?
  

4.4 Ruby annotation spanning: the 'ruby-span' property

- Similar to our previous comments on conformance, we think this spec
  should not say too much about XHTML applications' behavior.  In
  general, we're not fond of "special-casing" XHTML.

- What will happen if the value of attribute 'x' (which is supposed
  to be a number) is "0"?  Although the Ruby Annotation spec doesn't
  allow the value "0" for the rbspan attribute, this specification
  per se doesn't seem to prohibit that value for the 'ruby-span' property,
  and HTML 4 defined a funny behavior for similar attributes, namely
  rowspan and colspan attributes in table, when the value is "0".
  It would be better to clarify what is supposed to happen in this spec
  in that case.

    cf. http://www.w3.org/TR/html4/struct/tables.html#adef-rowspan

- The example in this section is not quite appropriate.  It sets
  the following style rule:

    myrtc  { display: ruby-text-container; }

  and an XML example uses two "myrtc" elements, yet this example
  doesn't specify any 'ruby-position' property.  Since the initial
  value for the 'ruby-position' property is "before", it's effectively
  the same as

    myrtc  { display: ruby-text-container; ruby-position: before; }

  for those two "myrtc" elements, and in section 4.1 the spec explicitly
  says as follows:  

    If two rtc elements are set with the same ruby-position value,
    (for example both 'before'), the relative position of the two
    elements is undefined. This setting should not be used.

  So the example should specify appropriate ruby-position for each
  "myrtc" element.

==========

Regards,
-- 
Masayasu Ishikawa / mimasa@w3.org
W3C - World Wide Web Consortium
for the HTML Working Group

Received on Wednesday, 27 November 2002 11:11:05 UTC