Re: [csswg-drafts] [css-flexbox-1] Merge section 9.2 with 9.3 (#11885)

> The reason we have them separated is that 9.2 is the steps necessary to determine the length of the lines (container main size), and 9.3 is what places items into those lines and sizes them.

[Point 3](https://drafts.csswg.org/css-flexbox-1/#algo-main-item), which constitutes the bulk of section 9.2, focuses on sizing the flex items themselves. It's unclear how this contributes to determining the length of the flex lines.  Is the actual goal to calculate the remaining free space available for distributing to the flex items in the main size ? Additionally, [point 2](https://drafts.csswg.org/css-flexbox-1/#algo-available) discusses cross space, but it’s not clear how that relates to line length, especially when the main size is in the inline direction and the cross size is in the block direction, which is arguably the most common interpretation of the spec.

Only [point 4](https://drafts.csswg.org/css-flexbox-1/#algo-main-container) appears to touch on sizing the flex line, assuming that the main size of the flex container determines the line length. However, that point requires further clarification on its own (see #11438).

That said, the header title feels somewhat backward. Since it appears that the flex line length is trivially equal to the container’s main size, I would expect the header to be “Main Size Determination.” That’s why I think it makes sense to merge section 9.2 with 9.3. The idea of line length could then be mentioned as a consequence of determining the main size, rather than being framed as the primary goal.

> Maybe it would help to caption these headers?

In general, I strongly support captioning headers, since headers alone rarely convey meaningful intent. As a principle, when describing algorithms, I favor minimal use of headers and more frequent use of explanatory "notes" between steps to clarify the purpose and logic behind them. That said, as mentioned above, I'm still unsure how the individual substeps in 9.2 contribute to the overarching goal of determining "line length".

 
> I'm somewhat ambivalent on whether to merge the sections in theory, but slightly biased towards not changing given how old the spec is--which increases the likelihood that people are referencing these sections by header name or number.

That's a fair concern. However, if these editorial changes were made in the context of Flexbox 2, I don't think backward compatibility would be an issue. In my experience, people usually quote the spec with a direct link, rather than by header alone. I’ve rarely, if ever seen someone refer to a specific section by name or number without also linking to it. So if we do want to merge headers, perhaps this is something best addressed in `css-flexbox-2` ?


Lastly, regarding "9.1. Initial Setup" I don't think a caption is necessary here. We can simply write:

"1.Determine the flex items, including anonymous items, as described in § 4 Flex Items."


-- 
GitHub Notification of comment by gitspeaks
Please view or discuss this issue at https://github.com/w3c/csswg-drafts/issues/11885#issuecomment-3016136855 using your GitHub account


-- 
Sent via github-notify-ml as configured in https://github.com/w3c/github-notify-ml-config

Received on Saturday, 28 June 2025 23:03:18 UTC