- From: Nat McCully via GitHub <sysbot+gh@w3.org>
- Date: Fri, 20 Aug 2021 17:32:44 +0000
- To: public-i18n-archive@w3.org
macnmm has just created a new issue for https://github.com/w3c/jlreq: == Proportional J text has spacing needs beyond JIS X 4051 mojikumi == JIS X 4051 describes conventional mojikumi tsume and aki (spacing adjustments) in terms of even fractions of 1 em, because the line box is usually exactly x em wide (c.f. [kihon hanmen](https://www.w3.org/TR/jlreq/#basic_templates_of_page_formats)), and so normally monospaced Japanese text with kinsoku clusters at the line ends or line starts only cause the lines to be short by 1/2 em or 1 em. To make the cluster fit by pushing in the line end, the mojikumi aki can be removed in even fractional (e.g. 1/2 em) increments, so the rules are in place to determine which character pairs should give up their space first, etc. In the case of full-justification and line width expansion, again the amount to be distributed among the character pairs can be expressed in even fractional em units, and so the differences in width for each pair between "desired" width and "maximum" width can also fit these even em units. These rules achieve two goals: One is to make the composer's job easier and more efficient by minimizing the places where adjustment must happen and in what amounts, the other is minimal disruption to the line's normal unadjusted width for maximum legibility and aesthetics. The issue becomes that when proportional type is introduced into the lines that necessarily create lines that are shorter than the line box width by amounts not easily divided among the 1/4 em or 1/2 em spacing around punctuation characters, or proportional punctuation glyphs whose aki space is not an even 1/2 em, you essentially have chaos in the available line width adjustments and no rules that deal with it nicely. The JIS X 4051 style of adjustment rules assume aki is always 1/2 em or 1/4 em or 1 em between punctuation and surrounding text, and has a dictated order of which aki is dealt with first, so the characters on the line return to a grid-like configuration as quickly and easily as possible. With proportional layout, the characters were never in that configuration to begin with, and so rules designed to preserve it are not suitable for such situations. In particular, the assumptions in JIS X 4051 need to be revisited, and new rules formed. 1. Aki amounts: not always 1/4 em or 1/2 em. Does this change priority rules at all? Does it change how adjustments should happen? 2. Text box width in even ems: The Kihon Hanmen of a embox grid seems well-suited for monospaced typography, but the needs of modern reflowable content and digital media seem to favor a much freer use of space when combined with proportional typography. This marriage of embox-based grids, embox-sized text or line boxes, and monospaced traditional Japanese layout, with the corresponding marriage of proportional layout within an arbitrary width, should be emphasized in the next JLReq. How to reflow content dynamically in either scenario should be explicitly described. The mixed use of arbitrary text box size with monospaced text, or, more rarely, proportional layout within an even embox-sized line box, can cause layout issues, bad adjustment decisions, and make beautiful line composition more difficult. 3. Aki contraction or expansion: Should proportional lines of type be adjusted differently or more evenly on a gradient in such cases? Now that adjustments can be done automatically in software, the prior historical considerations of efficiency in how and where adjustments are to be made by hand no longer apply. Please view or discuss this issue at https://github.com/w3c/jlreq/issues/296 using your GitHub account -- Sent via github-notify-ml as configured in https://github.com/w3c/github-notify-ml-config
Received on Friday, 20 August 2021 17:32:46 UTC