Whiteboard for January 2012 LCWD

This page is a status whiteboard for CSS3 Images. We needed a lot of whiteboard for the Disposition of Comments.

The DoC lists important messages on www-style in the course of discussion. This page is listing the actions the WG needs to take to resolve the issues. See these minutes for the discussion and resolution of these issues.

Directional Images

Defer directional images to L4 in order to address issue 37 (ltr/rtl annotations should be per-image() not per URL) and issue 41 (ltr/rtl annotations should be available via image-orientation)


Proposal A
Resolve on all of the following issues.
Proposal B
Drop object-fit and object-position.

I would advise on A, since there is only one significant issue. We have at least one implementation (Opera's; HP implements under the old names), and SVG would like these properties for mapping preserveAspectRatio into CSS. ~fantasai

Major Issues

Summaryobject-fit should not change size of content box
Option AKeep spec as-is
Option BRemove wording about resizing content box
NoteI'm ok with either proposal as long as dbaron agrees.
ActionThis issue requires a WG decision: A or B.

The intention of the text that's being proposed to remove is to solve the use case of scaling an image to cover or be contained by a particular 2D size (which is what 'object-fit' does) but also resize the content box to match the concrete object size (which 'object-fit' otherwise cannot do). The text does this by triggering the resize behavior on when 'width' and 'height' are both auto but the appropriate min/max constraints are set. A sample use case would be a photo album where each image must be sized to fit within a 100×100 square, but you want to put a box-shadow or a border on the image, not on the square.

Minor Issue

SummaryRemove clause allowing image-fit and image-position as aliases of object-fit/object-position
Note(Aliases were listed due to printer implementations.)
Option AKeep spec as-is.
Option BShift allowance to CSS Print Profile
Option CDrop allowance; such implementations are non-conforming.
ActionThis issue requires a WG decision: A, B, or C.

image() and Invalid Fragments

SummaryAllow fragment identifiers to be invalid and trigger image() fallback
OverviewKenny suggested treating media fragments as invalid images rather than requiring support in image(). This was rejected because the one major purpose of image() in this level is creating safe fallback behavior for authors using Media Fragments for spriting.
EditsHowever, for future media fragment extensions, we added a clause stating an unsupported media fragment syntax for a given image type makes the image invalid (triggering fallback to the next image in the image() list).
ActionThis issue resolution needs WG approval.

Inheritance of image-orientation

SummaryAllow image-orientation to inherit
EditsMade image-orientation inherit, which makes it more consistent with image-resolution and doesn't contradict the CSS Print Profile (which stated inheritance as “N/A”…)
ActionThis issue resolution needs WG approval.


Proposal A
Resolve on all of the following issues, aggressively solicit reviews from dbaron, roc, kenny, bzbarsky, and alexmog, and go to CR.
Proposal B
Defer element() to CSS4 Images.
Too many open issues and unreviewed significant changes.

NOTE: Tab has removed out-of-document element references, which were the source of some of the more significant issues.

Design-level Issues

Below is a summary of issues that should have explicit WG decisions:

SummaryGCPM element() and Images element() conflict
NoteSuggestion to rename Images element(), but no proposed resolution.
ActionHow does the WG want to resolve this conflict?
SummaryUse of 'bounding box' is undefined, should be 'border box'.
EditsAdded definition (new paragraph/list)
ActionDoes the WG approve of using the bounding box of the border image areas as the element() image bounds?
SummaryAllow image() to accept element() so that authors can specify fallbacks
EditsSyntax updated, defined when element() is invalid.
Action Does the WG approve of allowing element() in image() and defining an element() that {is not rendered and does not provide a paint source} to trigger image() fallback?
Summary Specify handling of varying-size pages
EditsDefined to align page content boxes by their start content edges before taking the bounding box.
ActionDoes the WG approve of this method of gluing pages together? Note that this definition may need to be reused for other things in the future.

Detail-level Issues

Below is a list of issues that don't require WG attention; but the WG should be aware they exist and in some cases the commenters need to reply and verify that the edits are correct:

SummaryPaint sources insufficiently defined
EditsAdded definition (new section)
SummaryCycle detection phrasing implies incremental algorithm
EditsAlgo reworded
NoteNo verification from commenter (dbaron).
SummaryCycle detection algo error
EditsAdded third bullet in cycle detection algorithm.
NoteNo verification from commenter.
SummaryLink to paint source definition for HTML
EditsLink updated to WHATWG HTML
SummarySpecify behavior when element() matches multiple elements
EditsSpecified to take first such element.
SummaryClarify “not rendered”
EditsDefinition added
SummaryClarify whether ancestor's perspective affects element()
NoteUnclear whether any edits were deemed necessary. No verification from commenter.
spec/css3-images/lc-2012.txt · Last modified: 2012/03/21 17:24 by fantasai
Recent changes RSS feed Valid XHTML 1.0 Valid CSS Driven by DokuWiki