IRC log of wai-wcag on 2003-06-11

Timestamps are in UTC.

13:59:57 [RRSAgent]
RRSAgent has joined #wai-wcag
14:00:02 [rscano]
hi Wendy :)
14:00:07 [wendy]
14:00:36 [rscano]
zakim, list conference
14:00:36 [Zakim]
I see Team_W3M()8:00AM, WAI_WCAG(techniques)10:00AM, DI_DIWG()10:00AM active and no others scheduled
14:01:03 [rscano]
zakim, this is WAI_WCAG(techniques)
14:01:03 [Zakim]
ok, rscano
14:01:11 [rscano]
zakim, who is on the phone?
14:01:11 [Zakim]
On the phone I see ??P6, ??P8
14:01:28 [Zakim]
14:01:50 [rscano]
zakim, who is speaking?
14:01:54 [Zakim]
14:02:01 [Zakim]
rscano, listening for 10 seconds I heard sound from the following: ??P6 (85%), ??P8 (5%), ??P9 (8%)
14:02:08 [rscano]
zakim, ??P6 is Roberto_Scano
14:02:09 [Zakim]
+Roberto_Scano; got it
14:02:11 [rscano]
14:02:11 [Michael]
Michael has joined #wai-wcag
14:02:26 [rscano]
zakim, ??P9 is Michael_Cooper
14:02:26 [Zakim]
+Michael_Cooper; got it
14:02:36 [Zakim]
14:02:41 [rscano]
who is P8?
14:03:06 [Michael]
zakim, I am Michael_Cooper
14:03:06 [Zakim]
ok, Michael, I now associate you with Michael_Cooper
14:03:07 [BBC]
BBC has joined #wai-wcag
14:03:14 [rscano]
zakim, who is here?
14:03:14 [Zakim]
On the phone I see Roberto_Scano, ??P8, Michael_Cooper, Wendy, ??P12
14:03:15 [Zakim]
On IRC I see BBC, Michael, RRSAgent, wendy, Zakim, rscano
14:03:25 [rscano]
P8 is Ben i think
14:03:48 [wendy]
zakim, ??P8 is Ben_Caldwell
14:03:48 [Zakim]
+Ben_Caldwell; got it
14:03:56 [wendy]
zakim, ??P12 is Chris_Ridpath
14:03:56 [Zakim]
+Chris_Ridpath; got it
14:04:01 [rscano]
zakim, who is here?
14:04:01 [Zakim]
On the phone I see Roberto_Scano, Ben_Caldwell, Michael_Cooper, Wendy, Chris_Ridpath
14:04:01 [wendy]
zakim, who's on the phone?
14:04:03 [Zakim]
On IRC I see BBC, Michael, RRSAgent, wendy, Zakim, rscano
14:04:04 [Zakim]
On the phone I see Roberto_Scano, Ben_Caldwell, Michael_Cooper, Wendy, Chris_Ridpath
14:04:05 [Al]
Al has joined #wai-wcag
14:05:12 [Al]
Zakim, is this WAI_WCAG?
14:05:12 [Zakim]
I don't understand your question, Al.
14:05:24 [wendy]
al, the code is 92248
14:05:28 [wendy]
(for zakim)
14:05:36 [rscano]
+1 617.761.6200 passcode 92248 ("WCAG-T")
14:06:21 [Al]
Zakim, please call AlGiman-home
14:06:21 [Zakim]
I am sorry, Al; I do not know a number for AlGiman-home
14:06:36 [Al]
Zakim, please call AlGilman-home
14:06:36 [Zakim]
ok, Al; the call is being made
14:06:37 [Zakim]
14:06:48 [rscano]
zakim, who is on the phone?
14:06:48 [Zakim]
On the phone I see Roberto_Scano, Ben_Caldwell, Michael_Cooper, Wendy, Chris_Ridpath, AlGilman (muted)
14:07:17 [Zakim]
14:07:38 [Al]
Zakim, call AlGilman-home, please
14:07:39 [Zakim]
ok, Al; the call is being made
14:07:41 [Zakim]
14:07:56 [Al]
I am back
14:08:23 [wendy]
23 techniques bugs have not been touched in the last week. want to make sure they are being handled.
14:08:44 [wendy]
wac has several new bugs, assign them?
14:09:05 [wendy]
action wac: change status on new bugs (assign them to me)
14:10:10 [wendy]
other work items from last week:
14:10:29 [wendy]
action bc: send info to richard
14:10:43 [rellero]
rellero has joined #wai-wcag
14:10:45 [wendy]
found weird diffs in browsers.
14:10:49 [rellero]
14:10:55 [rscano]
14:12:33 [wendy]
wac got all action items done except:
14:13:02 [wendy]
action wac: make sure the script (cleaned up) is available somewhere once it stabilizes - for use in the documents and as example of accessible script
14:13:17 [wendy]
action Wendy: finish moving HTML and CSS issues into Bugzilla
14:13:45 [wendy]
RRASAgent, replace action 2 with send Richard suggestions for accessibility of document and script
14:14:30 [wendy]
14:14:31 [wendy]
14:15:37 [wendy]
wac reports what steven (chair of html wg) said: there are better mechanisms to achieve the effect of layout tables. tables are for data.
14:16:04 [rscano]
for layout tables, a summary (i think) is not needed... so why don't put summary="*" so the user agent / AT knows that it is a layout table?
14:16:15 [wendy]
if xhtml 2 and forward will not support layout tables, could make hard and fast rule (xhtml 2 is a different technology).
14:16:52 [wendy]
table markup is used for layout, it contains graphical representation of relationships.
14:17:10 [wendy]
there is one pattern of relationsihps that we think of as data table.
14:17:23 [wendy]
making layouts, authors are using a more general pattern.
14:17:59 [wendy]
there are hints to the semantic links between the pieces.
14:18:18 [wendy]
need to understand the semantic structure of the layout tables before what we have requirements in xhtml2.
14:18:39 [wendy]
if not call them tables, then what call them? wysiwyg will have a tool to create something like a table.
14:18:43 [wendy]
what markup will it create?
14:19:07 [wendy]
(that's phase 1 - understand the uses of layout tables)
14:19:27 [wendy]
phase 2 - if you treat layout table as a scene graph there are 2 types of relationships. parts of wholes.
14:19:31 [wendy]
a row is an object.
14:19:47 [mcappelli]
mcappelli has joined #wai-wcag
14:19:49 [wendy]
row of headers is type description of the object (contains field names) on the subparts (column heads)
14:19:53 [rellero]
14:19:56 [mcappelli]
hi everybody
14:19:59 [rscano]
14:20:14 [wendy]
to understand cell need primary key and column head (tell what property, aspect is being described by the cell)
14:21:39 [wendy]
asst techs repeat some context as they read a table, but it varies.
14:21:57 [wendy]
they only tell qualifier for what changed.
14:22:30 [wendy]
if asst tech can build graph and understand relationships between cells can speak in way that works
14:22:43 [wendy]
we need to get relationships between elements captured so the AT can scrape it out
14:24:22 [wendy]
fedstats workshop on marking up tables
14:24:28 [wendy]
(need pointer)
14:24:49 [rellero]
In CSS3 a relational structure is foreseen among "id", so as to exceed the only current limit in the tableless layouts (columns are independent and have now different length according to the contents)?
14:27:24 [wendy]
techniques for html 4.01, xhtml 1, xhtml 2 could be very different, since diff technologies
14:27:43 [wendy]
so, what recommend say today?
14:27:44 [wendy]
default in AT is 1st col, 1st row
14:28:15 [wendy]
test how widely implemented th, scope, etc. are implemented.
14:28:42 [wendy]
using th instead of td for some cells is good practice, but need to get cells connected w/qualifiers.
14:29:08 [wendy]
proposals from last week:
14:29:15 [wendy]
th required for all cells that perform header role
14:29:26 [wendy]
if more complex structure, id, scope, etc.
14:29:39 [wendy]
data tables only those that contain header cells?
14:29:47 [wendy]
assuming working with data table, headers are required in that sit.
14:30:54 [wendy]
not th, but id/scope be required. since default handling of th is not enough.
14:31:26 [wendy]
while screen readers may default to 1st row/col. also want tables to work in other tools than screen readers (who have previous stated defaults)
14:31:37 [wendy]
header markup would help avoid misinterpretation. need to test this theory.
14:33:01 [wendy]
current proposal is "here are the features of HTML 4 and how to use them". is feature-driven not cost-driven for author.
14:34:33 [wendy]
is there a clear distinction between layout and data tables?
14:35:08 [wendy]
people seem to want clear rules. some may not like it, but most seem to want it.
14:35:23 [wendy]
thus, push for saying "don't put th in a layout table" seems to make things easier for everyone.
14:35:46 [wendy]
is there a strong reason why they would want to use th?
14:35:53 [wendy]
because of mixed-use tables.
14:36:04 [wendy]
partly data and partly layout. could stress to avoid doing that.
14:36:13 [wendy]
should use nested structures?
14:36:25 [wendy]
use css and tables.
14:36:30 [wendy]
but are css implementations ready?
14:36:45 [wendy]
still hearing that css is not a replacement for tables
14:37:34 [wendy]
it is possible, but it is a lot of work
14:37:56 [wendy]
question of: is the relationship outside of default reading order is important?
14:38:30 [wendy]
if a layout tables default reading order is ok, then stay away from other markup.
14:39:09 [wendy]
if a layout table did have relationships between cells, the table markup might be useful?
14:39:17 [wendy]
how do you represent that relationship?
14:39:27 [wendy]
can use id and headers, but can't describe it in detail.
14:41:20 [wendy]
jaws reads order in "headers" order. part of technique to put headers in order that want to be read.
14:42:38 [wendy]
if a table has a structure beyond layout but is not a header to cell relationship (part/whole, or generic relationship) then that is a semantic that might appear and could be worth dealing with.
14:43:12 [wendy]
work for sibling relationships?
14:44:49 [wendy]
if there is another cell that is required to understand the contents of this cell make the connection between those cells clear in the markup and let the assistive technology work out the reflow of the contents of those cells to give the user (of the AT) an experience that's easy to understand.
14:46:26 [wendy]
this rule could apply to any table, however people will want simple rule. i.e., "for data tables..." and for "layout tables"
14:46:34 [wendy]
could we have a hierarchy?
14:46:45 [wendy]
(of checklist items)
14:47:05 [wendy]
what should we tell the author vs what should we tell the UA/AT?
14:48:41 [wendy]
tables at either end of the spectrum (data <----> layout), here we can apply pretty clear/simple "rules", but it's those in the middle where we want to apply the top-down rule.
14:49:01 [wendy]
process of linearization means that semantic-free version must make sense.
14:49:22 [wendy]
if a linear order is enough, then leave to the UA.
14:49:59 [wendy]
in case where author has indicated layout, then push to discover what are the significant relationships to know as go through linear presentation.
14:50:23 [wendy]
requires knowledge of the relationships.
14:51:01 [wendy]
1. here's classical data table, do x. everything else is layout.
14:51:15 [wendy]
2. if obvious from reading order, don't need anything else. if not, need to do more.
14:51:44 [wendy]
tested sample tables where it does not have classical data table structure, but other structures that should be annotated? (headers in middle of table, etc.)
14:54:46 [wendy]
still not hearing an example of a visual relationship that needs to be represented using "data table markup"
14:55:21 [wendy]
the use of h1-hn, paragraph, in tables - did not expect (in design of relationship markup)
14:55:38 [wendy]
it makes those points navigable but not clear that makes their scope clear.
14:55:56 [wendy]
example: newspaper layout
14:56:31 [wendy]
full wide header block (masthead), left bar, right bar, multiple columns. cols might have headers. there is a colhead function.
14:58:20 [wendy]
collect a sample of layout tables and analyze.
15:00:49 [wendy]
within a markets table that is not marked up as data:
15:01:33 [wendy]
w/out support in Asst. Techs it is theoretical
15:03:29 [wendy]
nothing about block structure in the proposal
15:04:21 [wendy]
call out: layout table can contain data table, but data cannot contain layout
15:04:44 [wendy]
might need to use a nested table
15:05:04 [wendy]
using class or keyword in summary (related, but distinct)
15:05:14 [wendy]
class proposal grew out of summary proposal
15:05:40 [rscano]
using special charset like summary="*" ?
15:06:06 [rscano]
so if summary = "*" then skip because it is a layout table
15:06:46 [wendy]
people have been using summary="" - it's like null alt-text
15:07:07 [wendy]
at 1.0, suggested 3 techniques. 2 of which aren't viable in browsers.
15:07:43 [wendy]
method to track UA imps in techniques database
15:08:20 [wendy]
primary mandate? - create techniques that work today. but perhaps many places where we say "here's a future tech that we want techs to imp"
15:08:41 [wendy]
once supported, move into "today"
15:14:48 [wendy]
can't resolve until we do testing.
15:16:43 [wendy]
1. take a table and change markup to see how AT handles differently
15:16:55 [wendy]
2. talk with ATs to find out algorithms
15:17:24 [wendy]
web search w/a markup pattern
15:17:34 [wendy]
e.g., table w/colgroup
15:18:34 [wendy]
these are the types of tables to look at
15:23:10 [wendy]
mediametrix is now comscore:
15:23:15 [wendy]
1. how are tables being used
15:23:21 [wendy]
2. how are ATs handling tables
15:24:29 [wendy]
wac: interested in looking at sites
15:24:37 [wendy]
ag: interested in what the structure should be
15:24:49 [wendy]
mc: collect some examples and forward
15:24:59 [wendy]
bc: look at current support in AT
15:27:45 [wendy]
wac: play w/markup
15:27:52 [wendy]
assumption there is hybrid tables
15:29:17 [wendy]
wac: forward to list
15:29:56 [wendy]
plan to discuss this for perhaps 1/2 hour next week, no more.
15:30:43 [wendy]
next week:
15:30:46 [wendy]
1. tables 1/2 hour
15:30:49 [wendy]
2. schema
15:30:52 [wendy]
3. html techs draft
15:31:12 [wendy]
4. process for incorporating these discussions into techniques doc
15:31:15 [Al]
I won't be able to join next week [INCITS V2 F2F] but I'll try to look
15:31:23 [Al]
at sample pages.
15:32:12 [Al]
Al has left #wai-wcag
15:38:55 [rellero]
15:39:06 [rscano]
15:39:21 [mcappelli]
15:40:20 [mcappelli]
mcappelli has left #wai-wcag
15:40:31 [rscano]
rscano has left #wai-wcag
16:01:35 [wendy]
RRSAgent, bye
16:01:35 [RRSAgent]
I see 4 open action items:
16:01:35 [RRSAgent]
ACTION: wac to change status on new bugs (assign them to me) [1]
16:01:35 [RRSAgent]
recorded in
16:01:35 [RRSAgent]
ACTION: bc to send info to richard [2]
16:01:35 [RRSAgent]
recorded in
16:01:35 [RRSAgent]
ACTION: wac to make sure the script (cleaned up) is available somewhere once it stabilizes - for use in the documents and as example of accessible script [3]
16:01:35 [RRSAgent]
recorded in
16:01:35 [RRSAgent]
ACTION: Wendy to finish moving HTML and CSS issues into Bugzilla [4]
16:01:35 [RRSAgent]
recorded in