Restore scope section as issues.

Author: tabatkins

css-scroll-snap/Overview.bs

@@ -82,7 +82,6 @@ Use Cases {#use-cases}
 		}
 		li {
 			scroll-snap-align: edges;
-			scroll-snap-scope: finite;
 		}
 		</pre>
 
@@ -95,7 +94,6 @@ Use Cases {#use-cases}
 		}
 		.city {
 			scroll-snap-align: center;
-			scroll-snap-scope: finite;
 		}
 		</pre>
 </div>
@@ -190,10 +188,12 @@ On the children:
 			<td>n/a
 			<td>''scroll-snap-area: [ border-box | margin-box ] || <<length>>{1,4}''
 			<td>High priority
+		<!--
 		<tr>
 			<td>n/a
 			<td>''scroll-snap-scope: infinite | finite''
 			<td>Only ''finite'' is really necessary to handle all use cases, however UAs might want ''infinite'' as the initial value for performance reasons.
+		-->
 </table>
 
 Element-based Snapping {#element}
@@ -298,6 +298,69 @@ Scroll Snapping Alignment: the 'scroll-snap-align' property {#scroll-snap-align}
 	Issue: Think about how mandatory/finite would work when there's nothing within range to snap to.
 -->
 
+
+Scope of Snapping Influence {#scope}
+===========================
+
+	Issue: Current spec doesn't define how to select which snap-point to snap to.
+	See <a href="https://lists.w3.org/Archives/Public/www-style/2015Jul/0325.html">https://lists.w3.org/Archives/Public/www-style/2015Jul/0325.html</a>
+	for a proposal to ignore snap positions far outside the viewport.
+
+	Issue: UAs should be encouraged to ignore snap positions that require scrolling in two dimensions
+	when a one-dimensional scroll is triggered.
+
+	Issue: Define that snap-point selection is based on the final scroll position that the scroll physics would land the scroller in after a fling.
+
+<!--
+Scoping Snaplines to Visible Objects: the 'scroll-snap-scope' property {#scroll-snap-scope}
+---------------------------
+
+	<pre class="propdef">
+	Name: scroll-snap-scope
+	Value: infinite | finite
+	Initial: infinite
+	Applies to: all elements
+	Inherited: no
+	Computed value: as specified
+	Animatable: no
+	Media: visual
+	</pre>
+
+	When ''finite'' snapping is enabled,
+	the "gravitational field" of a snap	alignment is two-dimensional:
+	distance to the snap position is calculated for both dimensions at once.
+
+	In other words, if the snapping radius of influence is <var>r</var>,
+	in infinite	snapping the box snaps along the y axis
+	whenever it is within <var>r</var> of its snapped y position,
+	regardless of its x position.
+	But in finite snapping,
+	the box snaps along the y axis
+	whenever it is within <var>r</var> of its snapped position
+	in both dimensions.
+
+	<div class="example">
+		For example, a small box is snapped to the center of the viewport.
+		It only snaps whenever it is < <var>r</var> distance in any direction
+		from its snap position in both dimensions.
+		In other words, it snaps whenever sqrt(<var>d<sub>x</sub></var><sup>2</sup> + <var>d<sub>y</sub></var><sup>2</sup>) &le; <var>r</var>
+		for <var>d<sub>x</sub></var>,<var>d<sub>y</sub></var> as distance to the snapped position in the x and y dimensions, respectively.
+	</div>
+
+	<div class="example">
+		As another example, a small box is snapped to the edges of the viewport.
+		It only snaps whenever matching edges are within <var>r</var> of the respective viewport edges,
+		so e.g. whenever its top edge approaches the top of the viewport,
+		or its left edge approaches the left of the viewport;
+		but there is no snapping effect if those edges are > <var>r</var> outside the viewport.
+	</div>
+
+	Issue: This feature can be safely deferred to a future level, if necessary.
+	Alternately it can be dropped and ''finite'' snapping can be the default.
+	(We can't think of a use case for the infinite snapping model,
+	except perhaps UA performance.)
+-->
+
 Group-based Snapping {#group}
 ========================