Re: CJKフォントのpalt（詰め組）とkernの関係

ちょっと全部追いつくのが難しいですが、議論が深くなる前にゴールが合致していないので、まずゴールについて確認したいと思います。

フォントが存在するかどうか、という議論については、コードの互換性を壊す提案をしようとしているので、「存在する証明」ではなく、「存在しない証明」が必要だと思っています。これがないと、実装者から受け入れてもらうのは難しいのではないかと思いますが、これはまだできていませんよね？

次に、私が誤解しているかもしれませんが、村田さんの議論は、paltを持つ和文フォントが使われていたら、デフォルトでカーニングをオフにする、というところから始まっていますか？
欧文でもデフォルトのカーニングをオフにする、というのは、ユーザーからも実装者からもなかなか受け入れてもらい難いと思うのですが、私がなにか勘違いしていますか？


On Mon, Apr 17, 2023 at 6:35 AM Taro Yamamoto <tyamamot@adobe.com> wrote:

> Nat,
>
> Thank you for your prompt response. I’d like to add a minimum number of
> comments, and correct a few grammatical errors in my original message.
>
> ***
>
>
>
> I understand the ‘Latin-only kern’ mode works also for Japanese
> true-proportional fonts. Correct?
>
>    - Nat:* Latin only means we apply kerning and likely unnecessary palt
>    only to certain Unicodes in the non-CJK ranges that normally expect kerning
>    to be on by default, for any font.*
>
> [Taro: Ah, I understood. The apps. selectively apply both of the features
> to “Latin” characters in the ‘Latin’ ranges of code points. But I believe
> this is okay, as it won’t betray our expectations as a whole. . . .]
>
>
> As a Western proportional or Japanese true-proportional font lacks the
> ‘palt’ information usually, the ‘kern’ should work, if it has ‘kern’
> information. Correct?
>
>    - Nat: *yes  *
>
> [Taro: I’m relieved. Thanks.]
>
>
> But for a Japanese font without ‘palt’ information, neither the ‘palt’ nor
> the ‘kern’ feature work. Correct?
>
>    - Nat: *correct  *
>
>  [Taro: Fine. I corrected a double-negative error in my original
> sentence. 😉]
>
>
>
>    - *Nat: we generally do not check the font for the presence of a
>    feature to make decisions about other features. So, the engine will apply
>    palt if the user chooses "proportional widths". If the user chooses a
>    kerning mode, the engine applies palt and kern both. Engines that do not
>    apply palt whenever they apply kern are not correct according to the
>    standard and to our understanding of kern amounts set in the kern feature
>    for cjk glyph pairs. *
>
>
>
> Yes, but when a font lacks any ‘palt’ information, the ‘palt’ feature just
> doesn’t work, and the ‘kern’ feature is applied. Correct?
>
>    - Nat: correct
>
> [Taro: Okay. Dark clouds above us seem to be going away. . . .]
>
>
> And, as mainstream Japanese fonts with the alternative width capability
> always have ‘palt’ information, the ‘palt’ feature only or both the ‘palt’
> and ‘kern’ features should work always for those fonts. Right?
>
>    - Nat: *yes*
>
> [Taro: Good.]
>
> If so, the kind of fonts to which the ‘kern’ feature only is applied
> (without the ‘palt’ feature) is limited to Latin or Japanese
> true-proportional font. Right?
>
>    - Nat: *yes*
>
> [Taro: Great. Most of our potential points of concern seem to have been
> resolved, don’t they?]
>
> If so, I think the understanding mentioned by Mr. Murata is valid and
> suffices.
>
>
> (3) When a font specifies the 'kern' feature as well as the 'palt' feature
> for a given glyph, the abstract font engine shall not use the 'kern'
> feature for rendering this glyph without using the 'palt' feature as well.
>
>
>
> Font makers will need to just decide:
> *For Japanese fonts:*
>             (1) No ‘palt’, no’kern’
>             (2) ‘palt’ only
>
>             (3) both ‘palt’ and ‘kern’
>
> *For Latin or Japanese proportional fonts:*
>
> (4) No ‘palt’, No ‘kern’
> (5) ‘kern’ only
>
>
> I think the interaction between (3) above and these possible options on
> the font side seems reasonable and simple enough.
> What do you think?
>
>    - Nat: *seems fine to me  *
>
> [Taro: Glad to hear it. To celebrate our having reached this point, I’d
> like to quote Thomas Jefferson:
>
>
> *I never before knew the full value of trees. Under them I breakfast,
> dine, write, read and receive my company.*
>
>
> Well-designed fonts and good typography are like trees.
> Thank you again for your answers.
>
> Regards,
>
>
>
> --Taro
> ]
>