Re: [css-houdini-drafts] [text-layout] Houdini and text layout (#854)

Thank you @litherum and @AmeliaBR for sharing your thoughts. I agree that there are many aspects in text/inline layout, and opening all levels won't happen at once. It'll be a long process, adding capabilities gradually. Defining possible scopes, and which scopes should go earlier, is definitely the first topic I would like to discuss, and I appreciate you kicked off this discussion.

Region is a very interesting idea I've never thought about. I think Houdini is a good place to discuss about the capability when, IIRC, requiring JavaScript was one of controversial topics in CSS Regions discussion. I might want to flow not only text/inline boxes but also block boxes and other layout types into Regions, but I'm definitely interested in this topic.

Line breaking is a feature area we often hear people wants to have more control, especially in i18n, and one of my interests. It is not small though, given the amount of information UA needs to expose to the JavaScript, and its performance requirements is very high as @litherum pointed out. While I wish it to happen earlier, I also feel it needs a lot of design discussion and it may take longer than I would like.

Text shaping and positioning glyphs to me is even lower level than text layout. Again this is one of the things I have personal passion about, but the amount of information UA needs to expose is even more than line breaking, and it has a lot with Unicode and OpenType. At this point, I feel it'd probably be more work to define than line breaking.

One thing I find special in CSS text/inline layout is the line box. This type of the box does not exist originally but are generated during inline layout. IIUC this is the last one in the @AmeliaBR's list, allowing JavaScript to position text/inline boxes into a line box, then to position line boxes into the block flow or any other parent layout algorithm is one of the features we can open to authors. I guess this is probably easier to start then other scopes for its similarity to block layout API, and probably less hot from the performance perspective because it deals with boxes, not with characters. But easier to start does not necessarily mean to go first, if other scopes look more interesting to authors.

What about we try to have as long list of scopes as possible, and have a brainstorming session at Feb F2F?

cc @bfgeek @eaenet 

-- 
GitHub Notification of comment by kojiishi
Please view or discuss this issue at https://github.com/w3c/css-houdini-drafts/issues/854#issuecomment-459313436 using your GitHub account

Received on Thursday, 31 January 2019 11:38:19 UTC