Meeting Minutes	OpenOffice.org

Weekly Meeting Virtual User Interface (VUI) Team

Date	March 21^th 2002
Time	10:00h, MEST
Location	Cupertino, eham-02
Attendees	Christian Lippka (CL) Daniel Rentz (DR) Hans-Peter Burow (PB) Christian Jansen (CJ) Oliver Specht (OS) Frank Schönheit (FS)
Minute Taker	FS (next: PB, DR,GT, CL, OS, FS)
Distribution List	http://ui.openoffice.org/protocols

Table of Contents

1 Action Items

2 Round table

1 Action Items

Item	Responsible	Status
previous items
Create a concept of Common UI factory design	FS	stalled
Toolboxes in dialogs must get the correct image list set (standard/high contrast)	All ui developers	open¹
High Contrast mode (HC mode) for Listboxes, Treelistboxes and Browseboxes must be handled where used with bitmaps	All ui developers	open ¹
Calling SetModeBitmap for all dialog controls with bitmaps	All ui developers	open ¹
Check context menu handling	CJ	done ²
new items
check naming convention for high contrast bitmaps, check build process for high contrast image lists	OS	new
create Bug for UNO-Accessibility of BrowseBox	FS	new
create design document for Accessibility API of the BrowseBox	DR	new

Comments on Action Items

These items are waiting for the implementation of the interfaces to set bitmaps/image list for high contrast mode and the creation of high contrast ready bitmaps
Regharding the selection upon opening a context menu, there are different conventions in different products. The common denominator is that when opening a context menu via mouse, on an object which does not belong to the current selection, the selection is changed to this object. Differences are whether or not the original selection is restored after the menu is closed.
We think that as a general rule, we should change the OOo behaviour so that the selection follows the mouse click (which currently is not the case), but is not restored afterwards (due to the efforts this would imply especially for more complex scenes than a tree list box).

2 Round table

CJ told us that Stella nearly finished creation of the high contrast versions of all our office bitmaps.
It arose that it is yet unclear how these new bitmaps are to be use to generate image lists. Currently, the bitmap tool decides from the first two letters of the output name of a image list which input bitmaps to use - this would require new naming conventions for the high contrast images (currently they have a suffix "_h").
OS is to clarify this.
DR told there there still does not exist a formal bug for the UNO Accessibility of the BrowseBox. FS is to write one.
Given that there is the possibility that DR can't finish the Accessibility implementations for the BrowseBox, FS asked DR to create some specification document for the to-be-implemented API, to ease a possible later handoff.
There was a discussion about the cursor control in the BrowseBox (mainly in the database components). Currently, there are some inconsistencies between the BrowseBox, Calc Sheets and Tables in Writer. Some of the different behaviour does make sense, some perhaps doesn't.
The attendees agreed that a comprehensive solution would require extensive specification effort, and that it would not be worth it, as there are no real problems with the current behaviour.
PB told us of a performance problem in the Accessibility API of the ComboBox. As it was decided that a combo and list boxes expose all their entries (even the not visible ones) as children, we may have a problem with AT tools. For example, accessing the ComboBox in the Help component (with about 4000 entries) with the JavaMonkey takes a very long time which prevents reasonably working with it.
As soon as a product build is available with the API implementations in the combo box, PB will do some profiling to investigate the reason of the performance problem. It may be possible that the concept as such (expose all entries as objects) is the problem itself. Together with the various bridges involved, the problem may be inherent ....
FS asked whether or not the current toolbox design is final. He claimed that especially the visualization of "checked" toolbox slots is suboptimal. CJ told us that Stephan Schaefer is still not finished with the implementations, and especially the checked slots are to be improved.

Page 1 of 1