Minutes: January 8, 2014 SVG Accessibility Task Force

Web link:



Protocols and Formats Working Group Teleconference 09 Jan 2015

See also: IRC log

      Fred and Rich
         1. SVG mapping spec needs review
         2. Taxonomy Straw man walk through
         3. Group type or role
         4. Item role/type
         5. guide group
         6. annotation Group
         7. Scale Group
         8. region group
      Summary of Action Items

<trackbot> Date: 09 January 2015

<richardschwerdtfeger> meeting: W3C SVG Accessibility Task Force

<richardschwerdtfeger> chair: Fred

<fesch> scribe:Rich

<richardschwerdtfeger> scribenick: Rich

SVG mapping spec needs review


<fesch> Rich: Wants svg mapping spec reviewed, but wants a couple of weeks
to update before review

<fesch> Rich: provided background on spec he wants reviewed

<fesch> Doug: wants to ensure description is computed correctly and covers
all cases

Taxonomy Straw man walk through


<richardschwerdtfeger> Fred: I do want to start on the taxonomy that is in
the email in that link

<richardschwerdtfeger> Fred: I have 2 agenda items today which is to walk
through through the strawman as as starting point and see where we want to
go next. ..

<richardschwerdtfeger> Fred: we need to decide on the additional
taxonomies. We need to talk about data interaction. We need to get all of
us at a high level to agree on what the basic components of the taxonomy


<richardschwerdtfeger> Jason: there is a reference from the ACM journals
that describe graphics as to how they classify them.

<richardschwerdtfeger> Jason: if there is other research we need to look at
I am happy to work with them

<richardschwerdtfeger> Jason: anyone doing literature searches in this area
I am interested in talking to

<richardschwerdtfeger> doug: he is going to send the name of the
publication to the group as it is copyrighted

<richardschwerdtfeger> doug: we strongly avoid referencing things that is
not openly available to w3c spec.

<richardschwerdtfeger> doug: I won’t access things that are beyond the pay

<richardschwerdtfeger> fred: IBM has access to some of them.

<richardschwerdtfeger> jason: let’s take a look at the proposal that has
been put up.

<richardschwerdtfeger> fred: this taxonomy is grouped hirarchically

<richardschwerdtfeger> fred: the first role is a connector which seems to
map well to the SVG connector

<richardschwerdtfeger> fred: a leader line is something that links a label
to an object in a diagram. … like an eye chart.

<richardschwerdtfeger> doug: I agree that is not a connector

<richardschwerdtfeger> doug: when I discussed SVG connectors this was the

<richardschwerdtfeger> doug: I think we should have a connector role

<richardschwerdtfeger> doug: it should be implicit in the connector
element. In case someone wants to do something with a connector role
without using the connector spec. they should be able to do that.

<richardschwerdtfeger> fred: the one thing I am concerned about is spending
5 minutes on each item. I want tor try to get all of these today

<richardschwerdtfeger> fred: the next is a contour. It is a polygine. It is
a line of equal. One contour has the same value.

<richardschwerdtfeger> doug: for us to reasonably talk about them we need a
set of standard defintions.

<richardschwerdtfeger> fred: agree

<richardschwerdtfeger> fred: a contour is a topographic map line with
properties having the same value such as an elevation. They can be called
isobars as an example

<richardschwerdtfeger> fred: like temperature

<richardschwerdtfeger> event: I got this out of timelines

<richardschwerdtfeger> event: is something that is confined on a time line.

<richardschwerdtfeger> rich: an event needs to be in the context of
something so that it will make sense.

<richardschwerdtfeger> doug: for example an axis item or a legend item.

<richardschwerdtfeger> fred: for now I will leave it out in definitions as
it appears redundant

<richardschwerdtfeger> jason: it could be a subrole

<richardschwerdtfeger> fred: I have seen charts for something else. Such as
an industrial process where we are monitoring something.

<richardschwerdtfeger> jason: like between t1 and t2

<richardschwerdtfeger> fred: so where we are continuously montioring

<richardschwerdtfeger> jason: for someone who is not familiar with that
type of diagram they want to know how things are represented.

<richardschwerdtfeger> jason: we certainly need a system that in its firs
version or extensions that define the high level concepts where we can
define the high level meaning of things

<richardschwerdtfeger> jason: In an SVG diagram with formerly defined
semantics it is the user who defines which components of that content they
want to take

<richardschwerdtfeger> doug: I don’t think that we should take up the
visual aspect of this. That should be a separate topic.

<richardschwerdtfeger> jason: that is a good example of the issue

<richardschwerdtfeger> doug: I think the whole visual description is an
entirely separate topic

<richardschwerdtfeger> jason: the geometrical concepts can lead to this

<richardschwerdtfeger> doug: let’s try to defer for now

Group type or role

<richardschwerdtfeger> rich: we have group in the ARIA spec.

<richardschwerdtfeger> jason: perhaps datagroup role to be distinct from
the aria group role

Item role/type

<richardschwerdtfeger> fred: an item represents a single thing. It could be
a single point in a scatter plot. If you are familiar with a box with
whisker chart.

<richardschwerdtfeger> fred: it only means that it is a unique data item
vs. a graphical object

<richardschwerdtfeger> doug: when I did my chart I put axis and legend
items … i felt something like that to be an item but I looked at your item
as a datapoint

<richardschwerdtfeger> doug: it think we are conflating data points and

<richardschwerdtfeger> doug: do you think that having the distinction
between items as distinction between datapoints

<richardschwerdtfeger> fred: I think that data is different that points in
an axis

<richardschwerdtfeger> doug: yes

<richardschwerdtfeger> doug: I will let you finish. I think different types
of items are important

<richardschwerdtfeger> jason: we can introduce abstract roles where we want
to to classify concepts

guide group

<richardschwerdtfeger> fred: I took this out of statistical charts and maps

<richardschwerdtfeger> fred: thes are things that are not data directly

<richardschwerdtfeger> fred: there are two groups within

annotation Group

<richardschwerdtfeger> fred: I see these as labels

<richardschwerdtfeger> fred: under these we have label, note, and title

<richardschwerdtfeger> fred: the only distinction between notes and titles
is that notes don’t have particular meeting or location specific featues.

<richardschwerdtfeger> jason: how would a legend be handled?

<richardschwerdtfeger> fred: I have a scale group

Scale Group

<richardschwerdtfeger> fred: unde scape group I have axis, demnsion line,
graticule (they curve), grid, legend, and scale

<richardschwerdtfeger> fred: then we have a symble which is a wind
direction, … etcs.

<richardschwerdtfeger> cd/etcs/etc/

region group

<richardschwerdtfeger> this is different from guide group

<richardschwerdtfeger> doug: I don’t understand the distinction between
label and title

<richardschwerdtfeger> doug: most of the other stuff looks good

<richardschwerdtfeger> doug: rather than scale I would use facet

<richardschwerdtfeger> fred: they are different

<richardschwerdtfeger> doug: I would speparate them out and use facet

<richardschwerdtfeger> fred: Amelia likes dimensions

<richardschwerdtfeger> doug: facet is not in dimension space

<richardschwerdtfeger> fred: the only pushback on using facet vs. scale is
I would like to match domains people commonly use

<richardschwerdtfeger> doug: this is what I use for economics

<richardschwerdtfeger> jason: I think it is entirely reasonable but the
more high level comments we have we should describe them.

<richardschwerdtfeger> jason: I don’t do data visualization. so I need very
clear understanding of this

<richardschwerdtfeger> jason: we need real clarity

<richardschwerdtfeger> doug: I think that you will find that the literature
uses wildly different terms.

<richardschwerdtfeger> doug: I think you agree bar charts can be horizontal
and vertical

<richardschwerdtfeger> doug: some people insist bar charts can be vertical
where other people have a different trem for horizontal bar charts

<richardschwerdtfeger> doug: we are going to try to build a higher level

<richardschwerdtfeger> doug: we don’t want to slavishly follow existing

<fesch> rich: we have a note role in aria today

<fesch> rich: is it too early to try and write an aria spec?

<richardschwerdtfeger> rich: or write some definitions

<richardschwerdtfeger> fred: we can map these as a second step

<richardschwerdtfeger> rich: I am basically referring to an aria format

<richardschwerdtfeger> doug: fred came at it from a higher level of
abstraction and perhaps more exhaustive

<richardschwerdtfeger> doug: fred looked at it from a taxonomy so that we
understand the space more

<richardschwerdtfeger> doug: I don’t know that these are all roles

<richardschwerdtfeger> doug: I would like to see fred define his
terminology and then I define mine

<richardschwerdtfeger> doug: perhaps one more exercise in defining things
and then we go to a spec.

<richardschwerdtfeger> ACTION: Fred will create definitions for the next
meeting [recorded in

<trackbot> Created ACTION-1552 - Will create definitions for the next
meeting [on Fred Esch - due 2015-01-16].

<richardschwerdtfeger> ACTION: Doug create definitions for his taxonomy
[recorded in http://www.w3.org/2015/01/09-aria-minutes.html#action02]

<trackbot> Error finding 'Doug'. You can review and register nicknames at <

<richardschwerdtfeger> ACTION: shepazu create definitions for his taxonomy
[recorded in http://www.w3.org/2015/01/09-aria-minutes.html#action03]

<trackbot> Error finding 'shepazu'. You can review and register nicknames
at <https://www.w3.org/WAI/PF/Group/track/users>.

<richardschwerdtfeger> doug: I have one more question on symbols. You put
them outside. I think that is perfectly valid. I had thought that you were
talking about legend symbols

<richardschwerdtfeger> doug: we need to make a distinction between what you
have and what is in a legend

<richardschwerdtfeger> fred: for lack of something better I will write a

<richardschwerdtfeger> fred: this is where things are laid out on the paper

<richardschwerdtfeger> fred: I have seen these in blueprints and technical

<richardschwerdtfeger> fred: in some domains they appear to have meaning

<richardschwerdtfeger> doug: I had done interactive circuit diagrams and
facility diagrams

<richardschwerdtfeger> doug: I think we need to get the guys from Boeing

<richardschwerdtfeger> fred: that is it

<richardschwerdtfeger> fred: next time will have definitions and will go
through doug’s diagrams

<richardschwerdtfeger> rich: text at bottom:

<richardschwerdtfeger> Legend of Graphics Types

<richardschwerdtfeger> G Graph (statistical chart)

<richardschwerdtfeger> M Map

<richardschwerdtfeger> B Blueprint/Technical drawing

<richardschwerdtfeger> T Time line

<richardschwerdtfeger> doug: I think this task force should discuss
taxonomies. We should have one telecon talking about what the taxonomy is
for and the details of the spec.

<richardschwerdtfeger> doug: I don’t want the the svg working group to go
around in ignorance of what is being done and why

<richardschwerdtfeger> doug: I don’t want us isolated

<richardschwerdtfeger> doug: I would like us to discuss with them what aria
is and why

<richardschwerdtfeger> rich: we need to have a way to get at the specific

<richardschwerdtfeger> doug: for legend - we might want some thinking about
how a legend is differentiating different items on the chart. … for all the
blue items and all diamonds, etc.

<richardschwerdtfeger> doug: we should not avoid this

<richardschwerdtfeger> doug: I have thought a lot about the geometry and
how we should deal with it.

<richardschwerdtfeger> fred: i disagree with that. they need to have the
data value

<richardschwerdtfeger> jason: both the details and geometry is important

<richardschwerdtfeger> doug: I am not in complete agreement here

<richardschwerdtfeger> doug: before we do that I think we need to see how
it is exposed to AT.

<richardschwerdtfeger> fred: legends is a gnarley area for sure.

<richardschwerdtfeger> rich: this is the first time to ever tackle this

<richardschwerdtfeger> fred: I have a masters in mapping

Summary of Action Items
[NEW] ACTION: Doug create definitions for his taxonomy [recorded in
[NEW] ACTION: Fred will create definitions for the next meeting [recorded
in http://www.w3.org/2015/01/09-aria-minutes.html#action01]
[NEW] ACTION: shepazu create definitions for his taxonomy [recorded in

Rich Schwerdtfeger

Received on Friday, 9 January 2015 17:13:42 UTC