Alex, It's been a couple of weeks now that people have had a chance to
review these comments. Here are my responses.
1) Rename accessibleAt to cellAt: I'll keep the name as is since others
did not also request the change.
2) Add cellIndex back in (was childIndex): I'll leave it out. Noone
else asked for it.
3) Add method to take a header cell and return an array of related data
cells: I'll leave it out since others have not also requested this.
4) Limit header cells to being in the table: I believe this should be
implementation dependent.
5) Remove old rowHeader/columnHeader: I agree. This was the proposal.
6) select/unselectRow/Column: I agree these should mimic the
application's GUI behavior. I think that's what I've documented at
http://dev.linuxfoundation.org/~ptbrunet/ia2/api/AccessibleTable2.idl
I'm in the middle of migrating to a new machine, but later today I'll
refresh the IDL. I see I have some minor changes to make:
- remove some tab characters (replace with spaces)
- remove polish notation
Pete
--
Alexander Surkov wrote:
Hi. I have several points/questions.
1. Consider to rename accessibleAt to cellAt.
accessibleAt name says nothing until you look at method arguments. I
think cellAt is more intuitive name.
2. It's might be worth to keep cellIndex.
Cell index is unique numeric identifier of cell accessible. If AT
wants to navigate through all cells then it's easier to have one loop
to run through all cells, otherwise AT is forced to have nested loops
which can be not optimal if table has spanned cells or cell holes. If
you'll decide to save cell indexes then we should have two methods
like cellAtIndex and indexOfCell to avoid table-cell-index attribute
usage because it might be not performant (because of other attributes
calculation).
3. Consider to have reverse methods for row/columnHeaderCells.
Possibly AT may want to get data cells by header cell.
4. Should header cells be in scope of table or should it be
implementation dependent? I mean if I run through all cell accessibles
then should I meet all header cells among with data cells?
5. rowHeaders/columnHeader shouldn't be inherited from
IAccessibleTable interface. At least because they don't address some
complex cases and we didn't achieve unique opinion how to implement
them and moreover we decided to introduce new IAccessibleTable2
interface :)
6. I think selectRow and etc should allow client depended
implementation if AT vendors insist on it. But actually I would
appreciate personally if spreadsheet applications would work in
similar way (for sighted users and AT users).
Alex.
On Tue, Jul 14, 2009 at 9:07 AM, Pete Brunet<[email protected]> wrote:
Thanks Jamie, I updated the IDL at
http://dev.linuxfoundation.org/~ptbrunet/ia2/api/AccessibleTable2.idl
James Teh wrote:
Hi Pete,
All looks great. Just two small thoughts:
On 14/07/2009 2:12 AM, Pete Brunet wrote:
6) Redefinition of rowColumnExtentsAtIndex to use a reference to an
accessible instead of a cell index.
I don't think the "Index" suffix makes sense here anymore; i.e.
rowColumnExtentsAtIndex should be renamed to just rowColumnExtentsAt.
f) Pending discussion of whether the old rowHeader and columnHeader
should be removed.
I concur with the others that they should be removed for this new interface.
Jamie
--
Pete Brunet
a11ysoft - Accessibility Architecture and Development
(512) 238-6967
pete @ a11ysoft.com
http://www.a11ysoft.com/about/
http://www.linkedin.com/in/petebrunet
Ionosphere: WS4G
_______________________________________________
Accessibility-ia2 mailing list
[email protected]
https://lists.linux-foundation.org/mailman/listinfo/accessibility-ia2
_______________________________________________
Accessibility-ia2 mailing list
[email protected]
https://lists.linux-foundation.org/mailman/listinfo/accessibility-ia2
