Conversation
mbostock
left a comment
There was a problem hiding this comment.
I looked at doing title before, and I wouldn’t change title’s filter; the olympians scatterplot with info tip is a nice example of where you do want to filter by the title channel.
I think we can always change the ariaLabel filter in the future if we can think of a use case. Let’s start here.
The “filter iff the channel is not just an enhancement” rule covers this case. On the tip mark the title channel is necessary to display a (meaningful) tip. |
That would imply that the derived tip would filter more strictly than the original mark? Interesting. It might work. |
|
it seems to work in #1704! |
b7633b7 to
93e232f
Compare
…1704) * the title channel is not filtering the mark, unless it's a tip mark (any mark that depends on this channel must declare it) * rename to sparse title --------- Co-authored-by: Mike Bostock <[email protected]>
93e232f to
19ca6f0
Compare
* null filter on the href channel closes #1696 * the title channel is not filtering the mark, unless it's a tip mark (#1704) * the title channel is not filtering the mark, unless it's a tip mark (any mark that depends on this channel must declare it) * rename to sparse title --------- Co-authored-by: Mike Bostock <[email protected]> --------- Co-authored-by: Mike Bostock <[email protected]>
* null filter on the href channel closes observablehq#1696 * the title channel is not filtering the mark, unless it's a tip mark (observablehq#1704) * the title channel is not filtering the mark, unless it's a tip mark (any mark that depends on this channel must declare it) * rename to sparse title --------- Co-authored-by: Mike Bostock <[email protected]> --------- Co-authored-by: Mike Bostock <[email protected]>
closes #1696
I wonder if we should not extend this to any channel that “enhances” the mark but is not a visual channel. That is, title and ariaLabel? (The “use case” of using them to filter marks out seems far-fetched.)