Web and Software: August 22

Miscellaneous

• Approval of minutes from August 15, 2007.

Review of action items

1. IN PROCESS: Allen Hoffman will sync provision 2 with AV group definition by next week.
2. IN PROCESS: Katie Haritos-Shea will write a proposal for tactile and aural user preference settings.
   • Mailing list discussion started
   • Katie will repost proposal because most of the conversation on the list was off topic.
3. Judy Brewer to collect all clarification and rationale for the authoring tools provisions for the final TEITAC report. Not required prior to August 17, 2007.
4. IN PROCESS: Shannon Rapuano to look at "if" provisions to determine if we need an "else" provisionalso.
   • See Shannon's post, August 16, 2007
   • If platform does not require interoperability this pushes the product into being a closed product.
   • On a platform that does have this richness we need to find another way to do it.
   • If the AT does it all then you meet the functional performance criteria, but it should explain why and how they work with AT.
   • When is it a closed system such as a Kiosk?
   • Makes sense to either direct IT to interoperate with AT where it can, or make it clear that the platform does not provide what is needed.
   • The closed provision has been for closed functionality, but the definition did not get updated. It is closed functionality. The other thing had to do with the API, if there is no API, or an API that does not do all these things then how is the software supposed to do it. Even if the API does all these things it does not provide all the information and screen read can't do everything it needs to. What if there is no AT so doing this does nothing for using the product, because the platform is closed. There are a lot of issues around this that are unresolved.
   • Tabled discussion for now since its tied up with the bigger issue of AT interoperability.
5. IN PROCESS: Peter Korn to start discussion of "platform software" definition on mailing list.
   • On today's agenda
6. COMPLETE: Gregg Vanderheiden will take a look at the definitions for use of the words "Web content".
   • Used only in two notes in definitions. Could change to "Web page".
7. Gregg Vanderheiden will look at the latest language for "Web page" definition in WCAG 2.0 and post it to the list.
8. COMPLETE: Sean Hayes Sean Hayes will recommend a re-wording for the cognitive interaction recommendation with best practices removed and add it to the list.
   • See latest interaction guidelines provision
   • Agreement to leave it and let the editorial working group change it if they think its not good. This is just a recommendation.
   • Alex Li has a problem with part 2, sometimes you can not move backward. He wanted to restructure part 2 and 3, stated it may come up better then a recommendation.
   • Action: Alex Li to recommend rewording for 2nd and 3rd items in the "Interaction Guidelines" provision.
9. COMPLETE: Andi Snow-Weaver to include a note in functional criteria section for Word look up, test reading, and spelling assistance.
   • Michele will double check why this is not in the draft.
10. COMPLETE: Andi Snow-Weaver will reword language provisions and Alex Li will provide a note.
    • Alex has withdrawn the concern.
11. IN PROCESS: Andi Snow-Weaver will draft something about concerns for Authoring Tools provision 3 and 4 based on minutes from today.
    • Will probably changed wording to "must"
    • Placeholder for the concerns was included in Friday's draft. Looking for help from Sean Hayes and Andrew Kirkpatrick.
    • Discussed Sean's concern regarding the authoring tool definition and how broad it is. He has sent something out to the list on 8/16/2007 to narrow the scope of what the authoring tool is in these 2 provisions.
    • Preference to change the definition of Authoring tool to be more specific. At least in these two provisions.
    • Provision says additional requirement for Web content... which is a hold over, we intended to have something more generally applicable.
    • Concern with interoperability with 3rd party tools.
    • Compatibility and Interoperability and what that means... Need to further discuss what "compatibility with" means. The concern has to do with the definition of authoring tools being to broad for these provisions.
    • EWG draft is frozen. Going forward to next meeting the groups are all trying to fix problems and bugs and they will be passing resolutions. There is a link from the 8/17 draft. Fixes are put into the comment page and anyone can edit the that page. New decisions can also go there. New draft creation will depend on number of changes. It will include what has changed, outstanding issues and concerns.

Technical Topics

Definitions

• Platform software
  • The basic notion that Peter wanted to capture was something that is attempting to be a platform that isolates applications underneath, such as cross platform, flash, and java, that do not need to know about the platform it is written on.
  • At what point do a set of services isolate you from the underlying platform and then become a platform and have to meet this provision?
  • Definition may say "partially or completely isolate"... Web browser is mentioned in 3. Does a Web browser completely isolate... plug ins that draw from the browser and the system. Yet we are citing the browser as being a platform. We have things that don't isolate completely. Where a long the lines do we use that trigger?
  • A browser is a platform to HTML. A plug-in is machine code that is compiled to the machine architecture, coordinates and uses services of the browser. It also has direct connection to the underlying platform and it is not isolated.
  • It appears that things can be isolated sometimes and sometimes not, and can be partial at times. Not sure how we want to word these. What do we want to say and guarantee by doing this?
  • Sean had some difficultly with "can allow" isolation, which stops it from being a definition with any crispness. All libraries can give you some isolation. The only reason we are defining this is so we can define what needs to be included in the platform. He has posted a potential change to this.
  • Peter and Sean have been discussing what this has to say. Idea is when a platform is completely isolated it allows no possibility for the application to reach the underlying layer. If there is an escape then the application can provide the services, or use the escape to reach out to the OS. Java is self-contained but does allow this back door in some cases. Some may or may not allow access to the underlying system.
  • There are few instances where platform software completely isolates the application running on it. Given that is the reality and the relationship is shifting, we should aim to have a more flexible definition.
  • Peter's definition of a platform is okay if we change what the platform has to do.
  • Sean has suggested new wording. Continue to discuss and try to close by next week.

Platform Software - Accessibility Services

• Draft provision
• Alternate proposal to address Allen's concern that "other software" may be interpreted to apply to the platform only and not the applications running on the platform.
  • Platform software shall provide access to a set of services that enable assistive technologies to interact with user interfaces sufficient to enable compliance with the "AT interoperability" and "User Interface Components" provision.
  • Suggestion that we may want to consider making this IT, because this language says, shall provide access to a set of services it allows for the existence of an escape hatch.
  • Idea of the platform not providing access, the APIs are just a set of interfaces that the platform is intended to implement. Problem with the idea of the platform providing the service since it my be a broker between the application and the AT. If its a software running in the underlying service its not the platform that hooks it up to the AT. Can only be responsible for what is running on top of it.
  • Agreed to re-word as: Platform software shall provide access to a set of services that enable applications running on the platform to interact with other assistive technology sufficient to enable compliance with the "AT interoperability" and "User Interface Components" provisions.

Editorial review of use of terms

• Andi's post on August 15, 2007
• Proposals reflected in August 17th draft submitted to TEITAC
  • Application software has a placeholder in definitions

Content - used in non-text content provision

• Suggestion that we use "components" instead of "objects". Objects have a meaning in this definition, but in programming it may have a particular meaning. We do not use object in WCAG, because of the meaning in HTML.
• Another suggestion that we say "all non-text used as part of the user interface". Concern that it may appear that it this not applicable to documents.
• This is overlapping with the interoperability provision. Is this worth noting as a bullet that meeting the AT interoperability provision is a way of doing this.
  • Control input does refer to the "user interface components" provisions. There is some overlap and some instances where it's not provided programmatically.
  • Draft a note on the bottom that there is overlap, but one does not cover the other. Some strategies may address both this and the other AT interoperability provisions.
• Suggestion that "elements" may be better then "components".
• Object is actually the right term and applies to user interface and documents. Object is used in other provisions.
  • We should talk to people who write content to see if it makes sense to authors.
• Concern about this should only apply to non-text objects that are presented to the user.
• Agree to go with "non-text objects that are presented to the user"

Time-Limit

• Some people may develop the user interfaces and others then cause it to have a time limit, but may feel it doesn't apply to them. It would be better to say "product" rather then determine where the time limit comes from, which could complicate it.
  • Products and services are the terms we're using throughout the standards.
  • Subpart A uses products and services and if Web is not included then subpart A would need to be re-written. We need a global term that we can use.
• Maybe we can remove the term altogether. It should apply to everything, hardware, software and service. You can scope it in a way that whenever we can control it we manage the time limit. A system integrator should be sure the package does not violate the time limit restriction.
  • In the WCAG discussion if you are an author you don't always know the server has timed out. "Time limit that is set by the content" distinguishes that the requirement is only for what an author can and can not control.
    • In the WCAG environment, it applies to a more general population. But in the case of a federal agency, there is control of the server timeouts.
  • In government agencies we can say they have control, but individual agencies may have very little control. So maybe having the provision that "set by..." is important. It also applies to copiers etc. Suggest "by the product including content..." as the language. Maybe look at the EWG activity to see how the word product is being used.
• Suggestion we say "set by the product" or "set by the product or content".
• Needs to be scoped so that it only applies to time limits that affect the user. The purpose is to give the user the ability to control the situation when they need to control it. Some time limits may not apply - only the ones where a user has a problem. Content is the right term.
  • "time limit that is set by the content", or
  • "time limit that is presented to the user"
• Concern that definition of content applies to almost everything - may have the wrong definition then.
  • some see software as content, others disagree.

New Action Items

• Alex Li to recommend rewording of interaction guidelines provision part 2 and 3.

Attendance

1. Tom Brett
2. Sharon Snider - IBM
3. Andi Snow-Weaver - IBM
4. Michele Budris, Sun Microsystems
5. Bruce Bailey (Access Board)
6. GreggVan
7. Eric Damery, Freedom Scientific
8. Alex Li (SAP)
9. Shannon Rapuano - IBM
10. Peter Korn, Sun Microsystems, Inc.
11. sean hayes (Microsoft)
12. Mary Beth Janes (Apple)
13. Peter Wallack (Oracle)
14. Jim Allan (w3C)
15. Angela Hooker
16. William Loughborough
17. Jim Elekes (Access Board)
18. Dana Louise Simberkoff-HiSoftware
19. terry weaver