[Accessibility] Draft minutes for 10/22/03

John Goldthwaite john.goldthwaite at catea.org
Wed Oct 22 14:07:44 PDT 2003


Meeting minutes 10/22/03
A meeting of the Accessibility workgroup was held by teleconference on
Wednesday, Oct. 22, 2003, at 18:00 UTC

Attending:
Stuart Anderson, LSB specification for FSG
Doug Beattie
Jonathan Blanford
John Goldthwaite
Bill Haneman
Randy Horwitz
Peter Korn
Janina Sajka
Gunnar Schmidt
Sharon Snider
Matthew Wilcox
Allen Wilson

Action Items:
Gunnar- contacting Trolltech about participating in the AT-SPI workgroup.
Janina =96 complete and circulate the press release

Minutes for last week approved with changes noted in emails.

Janina has a draft of the announcement.  Target release date Nov. 3
Rich Schwerdtfeger is developing a statement from IBM.  Bill has one from
Gnome in process.
Gunnar forwarded the following statement approved by KDE

<statement>
Although there are many ongoing efforts to help persons with disabilities
use Linux systems, most Linux GUIs are still inaccessible to blind persons
and present many obstacles to people who cannot operate a standard keyboard
or mouse.

We do not want to support this social exclusion but rather see the support
for handicapped persons as an important feature for every software. The KDE
accessibility team knows about the problem and they are working on improving
the accessibility of KDE.

In order to get a completely accessible system, applications that use other
toolkits need to be accessible as well. In this sense KDE based assistive
technologies need to interoperate well with applications using other
toolkits and vice versa. We appreciate the foundation of the FSG
Accessibility Workgroup and are very pleased to announce our participation.
<statement>

Stuart Anderson gave an overview of specification documentation.
Specs written in Docbook =96 can derive pdf, rtf,
Documents stored in a gforge process.  Can use it to produce a list of
changes between versions.  Have scripts to do =85
Things such as APIs would be pulled into the LSB in the future so it would
be good to have them in spec format.  Language is formal, standard format.
Other components such as input methods would be things that an application
would not talk to directly would be put in another area (level above).

Bill =96 what do you do when referencing other specifications?
Stuart- try to be as specific as possible, Spec. name, chapter, page.  You
can use the cross reference feature of Docbook.   Don=92t include the text =
of
the other spec. e.g. , =93it behaves as described in the xx specification on
page yy=94   Some standards are more rigorously published than others.  Best
if they are formally approved and published so there is a controlled
version.   C++ is updated asynchronously, now have timestamps to make them
easier to track.

Doug =96 putting SPI into a database, how do you model the database.
Stuart can provide a database template. Or we could start with a copy of the
LSB database and add our material. He runs the database for LSB in mySQL.
Bill- should be able to reference a stable xkb stack from the X site.  There
is more than just access in xkb so it would be a subset.
Stuart-
Insuring information is functions?  Distributions out there whose xkb ;
Need comformance testing.

Bill =96 we=92ve said we want to do on the keyboard in year one.  Is that
possible if there is some clean up that needs to be done on the whole spec?
Stuart =96 Could be done in a year.  Will need to create test as well.
-	clean up - few weeks
-	verify against other implementations and build test suite.

Bill- we had hoped to do the conformance testing later and just do the spec.
first.
Doug- having the tests are part of getting the standard accepted
Jonathan- how detailed does it have to be, can it be behavioral?
Stuart- fully automatic is preferred.  Motif is automatic but it tests only
the widgets and not the keyboard.
Who develops the tests?
Someone has paid for the test development and donated it.

Bill- there are two parts, the AT-SPI part and the activeX features that are
exposed to the user.  The latter will have to be functional.  Some will be
automatic and some will be check lists.
Stuart- Some things will fit into an automated methods and others just won=
=92
t.
Jonathan- at some point the LSB will want it so we need we should keep it in
mind.
Bill- On ABI part, are there any goals associated with ABI versioning-
Stuart- in series 1.1 .. 1.2 we make the new versions backward compatible
( don=92t break features written for the 1.1 version).   With major 1.0 to
2.0, there could be incompatibility.   Don=92t do a final approval until we
have a couple of implementations that can be tested.   Want to have two
major revisions to be sure we have everything in the spec.
Bill =96we should have a couple of implementations of the ABI if we include
Java, although the later is not open.

Item 2 =96 getting groups going:  AT-SPI, xkb, shared I/O

AT-SPI =96   Bill, Jonathan, Gunnar
Xkb =96   Bill, Earl Johnson, Gunnar
Shared I/O - Kirk, Mario, Willy =96

Jonathan nominates Bill to chair AT-SPI since no one else knows AT-SPI well
enough.

Gunnar =96 we should invite someone from Trolltech to join the AT-SPI.  They
are modifying their access infrastructure so they can create a bridge to
AT-SPI.
Janina- we need to cover all the communities of interest especially
implementers
Doug =96 Trace center?  Would they want to contribute some

Item 3 =96 press release draft will be distributed shortly.
We=92re not going to make Enterprise Linux conference.  We will try to get =
it
approved by November 3.

Our next meeting is Wednesday, Oct. 29 at 18:00 UTC



John Goldthwaite
Center for Assistive Technology and Environmental Access, Georgia Tech
john.goldthwaite at catea.org
-------------- next part --------------
Meeting minutes 10/22/03
A meeting of the Accessibility workgroup was held by teleconference on Wedn=
esday, Oct. 22, 2003, at 18:00 UTC

Attending: =

Stuart Anderson, LSB specification for FSG
Doug Beattie
Jonathan Blanford
John Goldthwaite
Bill Haneman
Randy Horwitz
Peter Korn
Janina Sajka
Gunnar Schmidt
Sharon Snider
Matthew Wilcox
Allen Wilson

Action Items:
Gunnar- contacting Trolltech about participating in the AT-SPI workgroup.
Janina - complete and circulate the press release

Minutes for last week approved with changes noted in emails. =


Janina has a draft of the announcement.  Target release date Nov. 3
Rich Schwerdtfeger is developing a statement from IBM.  Bill has one from G=
nome in process.
Gunnar forwarded the following statement approved by KDE

<statement>
Although there are many ongoing efforts to help persons with disabilities u=
se Linux systems, most Linux GUIs are still inaccessible to blind persons a=
nd present many obstacles to people who cannot operate a standard keyboard =
or mouse.

We do not want to support this social exclusion but rather see the support =
for handicapped persons as an important feature for every software. The KDE=
 accessibility team knows about the problem and they are working on improvi=
ng the accessibility of KDE.

In order to get a completely accessible system, applications that use other=
 toolkits need to be accessible as well. In this sense KDE based assistive =
technologies need to interoperate well with applications using other toolki=
ts and vice versa. We appreciate the foundation of the FSG Accessibility Wo=
rkgroup and are very pleased to announce our participation.
<statement>

Stuart Anderson gave an overview of specification documentation.
Specs written in Docbook - can derive pdf, rtf, =

Documents stored in a gforge process.  Can use it to produce a list of chan=
ges between versions.  Have scripts to do ...
Things such as APIs would be pulled into the LSB in the future so it would =
be good to have them in spec format.  Language is formal, standard format. =
 Other components such as input methods would be things that an application=
 would not talk to directly would be put in another area (level above).
   =

Bill - what do you do when referencing other specifications?
Stuart- try to be as specific as possible, Spec. name, chapter, page.  You =
can use the cross reference feature of Docbook.   Don't include the text of=
 the other spec. e.g. , "it behaves as described in the xx specification on=
 page yy"   Some standards are more rigorously published than others.  Best=
 if they are formally approved and published so there is a controlled versi=
on.   C++ is updated asynchronously, now have timestamps to make them easie=
r to track.

Doug - putting SPI into a database, how do you model the database.  =

Stuart can provide a database template. Or we could start with a copy of th=
e LSB database and add our material. He runs the database for LSB in mySQL.=
  =

Bill- should be able to reference a stable xkb stack from the X site.  Ther=
e is more than just access in xkb so it would be a subset.  =

Stuart- =

Insuring information is functions?  Distributions out there whose xkb ;  =

Need comformance testing.  =


Bill - we've said we want to do on the keyboard in year one.  Is that possi=
ble if there is some clean up that needs to be done on the whole spec?
Stuart - Could be done in a year.  Will need to create test as well.  =

- clean up - few weeks
- verify against other implementations and build test suite.

Bill- we had hoped to do the conformance testing later and just do the spec=
. first.
Doug- having the tests are part of getting the standard accepted
Jonathan- how detailed does it have to be, can it be behavioral?
Stuart- fully automatic is preferred.  Motif is automatic but it tests only=
 the widgets and not the keyboard.  =

Who develops the tests? =

Someone has paid for the test development and donated it.

Bill- there are two parts, the AT-SPI part and the activeX features that ar=
e exposed to the user.  The latter will have to be functional.  Some will b=
e automatic and some will be check lists.
Stuart- Some things will fit into an automated methods and others just won'=
t. =

Jonathan- at some point the LSB will want it so we need we should keep it i=
n mind.
Bill- On ABI part, are there any goals associated with ABI versioning-
Stuart- in series 1.1 .. 1.2 we make the new versions backward compatible (=
 don't break features written for the 1.1 version).   With major 1.0 to 2.0=
, there could be incompatibility.   Don't do a final approval until we have=
 a couple of implementations that can be tested.   Want to have two major r=
evisions to be sure we have everything in the spec.
Bill -we should have a couple of implementations of the ABI if we include J=
ava, although the later is not open.

Item 2 - getting groups going:  AT-SPI, xkb, shared I/O

AT-SPI -   Bill, Jonathan, Gunnar
Xkb -   Bill, Earl Johnson, Gunnar
Shared I/O - Kirk, Mario, Willy - =


Jonathan nominates Bill to chair AT-SPI since no one else knows AT-SPI well=
 enough.
  =

Gunnar - we should invite someone from Trolltech to join the AT-SPI.  They =
are modifying their access infrastructure so they can create a bridge to AT=
-SPI.  =

Janina- we need to cover all the communities of interest especially impleme=
nters
Doug - Trace center?  Would they want to contribute some =


Item 3 - press release draft will be distributed shortly.
We're not going to make Enterprise Linux conference.  We will try to get it=
 approved by November 3.  =


Our next meeting is Wednesday, Oct. 29 at 18:00 UTC




More information about the Accessibility mailing list