JLReq TF meeting notes 2019 December 13th

JLReq TF meeting notes 2019 December 13th

Attendee
–––––––––––––––––––––
Kobayashi, Tatsuo
Kobayashi, Toshi (Bin-sensei)
Murata, Makoto
McCully, Nat (online)
Murakami, Shinyu
Shimono, Atsushi
Tajima, Jun
Takase, Hiroshi
Kida, Yasuo

Meeting notes
–––––––––––––––––––––

JLReq updates
JLReq update's development and tests are now complete! (except the change log #149)
Remaining issues mentioned:
Default language should be set to English (#126)
U+200B garbage here and there → agreed to leave them alone as long as they do no harm.
Todo: Shimono-san will initiate the process for making it public.

GapAnalysis tests
Tajima-san, Shimono-san and Takasen-san are making good progress on text-spacing & text-indent properties*. Temporary workspace on a forked i18n-test repository.
(* Of 21 property tests that JAGAT developed, 10 already exist in i18n-test, 7 are for paged media, 2 require developing a font with specific properties and right now they are beyond our ability (some discussions below), and 2 are ones we can work on. They are text-spacing and text-indent. c.f. URL)

Brainstorming on testing layouts
Integrated tests
Adobe is working on tests that compare results
It would be worth proposing samples of good layout from the task force.
There are many house rules and no single “right” answer.
Pixel by pixel comparison does not make sense.
JAGAT test suites uses human eyes to check if the layout is done correctly.
There are properties that should result in single right answer, and there are others that leaves flexibility.
If the tests are done on individual properties (rather integrated one), the complexity might be easier to handle.
Paged media
You can see results on paged environment by looking at the “print preview” on various browsers. Chrome has a switch to emulate the behavior.
One way might be to do the tests on applications supporting the CSS Paged Media (Vivliostyle is one such example).
Fonts
Testing properties that require font features need fonts supporting these specific properties.
Can’t we just use fonts on each platform?
If we had free fonts in the industry standard Adobe-Japan1-6, we can depend on such fonts to develop tests. Unfortunately there aren’t such.
Kobayashi-san promised that he will take this issue to “文字情報技術促進協議会 (CITPC)” at which he is the chairman, and will evaluate the feasibility of developing fonts in Adobe-Japan1-7. It enables people to depend on such standard fonts for testing font related CSS properties. This is super!
Kida mentioned he hopes such fonts be free of charge for tests.

Page 2020
Page 2020 conference is on 2/7. The staff wants update from JLReq TF. Kida will prepare a 15 min presentation.

Github is not a good tool for non-computer-experts
It is becoming apparent that github is not necessarily a good tool for discussions that involve layout experts who are typically not savvy on computers. Right now such discussions are done on ad-hoc way on email including people who are relevant. When agreements are reached they are posted to the github with background.
Do we have a better solution than ad-hoc email threads? This was recognized as a potential issue between Richard, Kida, etc. when we were preparing to setup the task force. Kida will revisit it with Richard. Kobayashi-san wants it be with closed member to keep the discussions focused.

Gap Analysis: Review of CSS definition status
We’ve been reviewing JLReq sections to check if CSS properties are defined for requested features. The first pass was completed this time.

We want to do a second pass to make sure the data was gathered in a consistent manner, to separate cases between non-paged and paged media, and if the feature has impact to the readability.

4.2.2 Note Numbers <https://www.w3.org/TR/jlreq/#note_numbers> and 4.2.3 The Processing of the Reference Mark <https://www.w3.org/TR/jlreq/#the_processing_of_the_reference_mark>
× No corresponding CSS. Medium priority.
It is possible to imitate the layout, by using inline block, ruby, display property, etc.. Has issue with positioning. Making sure the number and the character / word to appear in the same line is important but not possible.
There is a draft specification for paged media - generated content for paged media <https://www.w3.org/TR/css-gcpm-3/>.
Still, fine control of positioning is not possible. This is a generic issue that appears with various properties.
TODO: We want to review generic issues after we’ve done reviewing JLReq sections.
Casted to JWT
4.2.4 Processing of Endnotes in Vertical Writing Mode or Horizontal Writing Mode <https://www.w3.org/TR/jlreq/#processing_of_endnotes_in_vertical_writing_mode_or_horizontal_writing_mode>
Paged media only
Low priority
4.2.5 Processing of Footnotes in Horizontal Writing Mode <https://www.w3.org/TR/jlreq/#processing_of_footnotes_in_horizontal_writing_mode>
△ The proposed draft of generated content <https://www.w3.org/TR/css-gcpm-3/> does not support the case where footnotes are set continuously following the base text. Such case is low priority.
4.2.6 Processing of Sidenotes in Vertical Writing Mode <https://www.w3.org/TR/jlreq/#processing_of_sidenote_in_vertical_writing_mode>
△ The proposed draft of generated content <https://www.w3.org/TR/css-gcpm-3/> does not support the case where sidenotes are positioned at the left side of a spread. Such case is low priority.
4.2.7 Processing of Headnotes (in Vertical Writing Mode), Footnotes (in Vertical Writing Mode) and Sidenotes (in Horizontal Writing Mode) <https://www.w3.org/TR/jlreq/#processing_of_headnote_footnote_and_sidenote>, i.e. parallel notes.
○ Possible even with non-paged media.
Future 4.57 possible with both sides ruby.
4.3 Positioning of Illustrations <https://www.w3.org/TR/jlreq/#positioning_of_illustrations>
△ It is not possible to avoid very short lines when the image is large. Also, when the remaining space is short and not a multiple of fullwidth, the space between characters become large (how we make sure the line length is multiple of fullwidth?). Solving these limitations is high priority.
Page floats possible with the Page Floats <https://www.w3.org/TR/css-page-floats-3/> draft.
Paged media: △ Basic positioning possible. There are limitations such as positioning against a spread.
4.4. Processing of Tables <https://www.w3.org/TR/jlreq/#processing_of_tables>
Most issues are common to 4.3 Positioning of Illustrations <https://www.w3.org/TR/jlreq/#positioning_of_illustrations>.
It is possible to fix the header position relative to the screen by position: sticky <https://www.w3.org/TR/css-position-3/#sticky-pos>.
Paged media: It is possible to repeat headers across pages (CSS Tables. Working draft. Section 6.2. <https://www.w3.org/TR/css-tables-3/#repeated-headers>)
× Figure 4.109, spreading a table in a spread, not possible.
4.5 Block Direction Setting Process of Lines, Paragraphs etc. <https://www.w3.org/TR/jlreq/#block_direction_setting_process_of_lines_paragraphs>
× Not possible to fix the line positions. High priority
Line Grid <https://www.w3.org/TR/css-line-grid-1/> draft is trying to achieve this with Rhythmic Sizing <https://www.w3.org/TR/css-rhythm-1/>.
4.5.2 Processing of Spaces between Paragraphs <https://www.w3.org/TR/jlreq/#processing_of_spaces_between_paragraphs>
○ margin property, with line unit “lh”.
Figure 4.124, not possible.
Paged media: × Justification of paragraphs and other blocks in the direction of line progress. Low priority.
//