Re: [css-line-grid][css-books] opt-in or opt-out? (was one property or more?

On Oct 3, 2014, at 10:10, Alan Stearns <> wrote:

>>> [1]
>> Not sure which picture in the blog post you think is common
> Either of the last two are valid ways to set headings in Latin. What you
> need to avoid is line snapping everything [4].

Agreed, but that doesn’t explain which is fewer, or whether opt-in or opt-out is better strategy. I agree with Håkon that by default all to snap and author specify what not to is better.

>> having blocks align to the grid looks natural to me, and matches to what
>> East Asians do[2]. If that’s the case, wouldn’t it be more appropriate to
>> say there are more elements to snap than not to?
> The more or less elements argument was applied to the use of line snapping
> - both drafts use a separate property for box/block snapping. So adding
> more box-snap elements isn’t that relevant to the question of whether we
> should have one or two properties for establishing a grid and turning on
> line snapping.

You’re right, I was discussing whether to opt-in or opt-out, so changed the subject.

>> I can’t really speak of Latin typography better than you, but please keep
>> in mind that, at least in East Asia, authors would want almost every
>> element to snap, except headings and pictures as blocks to snap (see Fig.
>> 4.20 for when headings are of multiple lines[3].)
>> [2]
>> [3]
> Right - Fig 4.20 shows a heading that box snaps and deliberately does
> *not* line snap. That’s quite similar to the examples I give in my blog
> post, though the box snap setting used will likely be different for
> different scripts.

As Christoph said, I’m not convinced that only few elements snap. I don’t want authors to make a list of “p, div, dl, dd, dt, ol, ul, li, …” in every styesheet.

One possible idea is to make everything snap by default, with:
* line-snap if the computed line-height is the same as the line-height of the element that created the grid, and
* box-snap otherwise.
Authors can further specify specific elements to behave differently, but I think that’d be a reasonable default for most of daily use cases.

Does this suffice both of requirements?


Received on Friday, 3 October 2014 06:10:10 UTC