Re: [css-line-grid] Are named grids required?

On 2/26/14, 1:21 PM, "Tab Atkins Jr." <jackalmage@gmail.com> wrote:

>On Tue, Feb 25, 2014 at 10:10 PM, fantasai
><fantasai.lists@inkedblade.net> wrote:
>> On 02/25/2014 04:44 PM, Alan Stearns wrote:
>>> I've been thinking about how we might simplify the initial version of a
>>> baseline grid feature while still allowing future extensions.
>>>
>>> One thing that the current draft and your proposal both have are named
>>> grids. I'm not certain named grids are required for the largest set of
>>>use
>>> cases. In most cases I can think of, there is only a single grid. And
>>>when
>>> you do need more than one grid, the use of each grid is usually
>>> constrained to separate parts of the tree (elements in a single parent
>>>do
>>> not snap to different grids).
>>>
>>> What if we only allowed unnamed baseline grids to start, but left the
>>> syntax open to allow for named grids in the future? The first
>>>iteration of
>>> the line-grid property would only allow an element to establish a new
>>> baseline grid:
>>>
>>> line-grid: auto | new
>>>
>>> where auto (or none?) is the initial value that doesn't establish a
>>>grid,
>>> and new establishes a new baseline grid that all of the element's
>>>children
>>> can use to snap to. Later on, if we find a need for named grids we
>>>could
>>> extend the syntax to:
>>>
>>> line-grid: auto | new [named <ident>]
>>>
>>> The line-snap property would start out just snapping the dominant
>>>baseline
>>> to whatever grid has been established for the element - the closest
>>>parent
>>> with 'line-grid: new' or a default grid from the root element. When
>>>and if
>>> named grids were added, then we could also add a keyword to the
>>>line-snap
>>> property to pick a named grid instead (perhaps 'from <ident>'?)
>>
>>
>> I was actually pondering the same thing, with the keywords
>>   line-grid: match-parent | create
>> :)
>>
>> I'm happy to update the spec as such, but thoughts on naming?
>> (Also, comments from other people?)
>
>I like "match-parent", but is it useful to allow something to opt out
>of the line-grid without establishing a new one?  (That is, an
>explicit "none" value?)

My thought has been that line-grid only determines whether an element
establishes a new line grid or inherits the existing grid from a parent.
It’s the line-snap (and box-snap or block-snap?) properties that allow for
opting in or out of using the grid for positioning boxes.

So while the initial value of line-grid needs to allow for using a
pre-existing grid, I’m not sure there’s a use for turning off the grid
from this property.

Thanks,

Alan

Received on Wednesday, 26 February 2014 21:32:20 UTC