29 May 2003 - WCAG WG Teleconference Minutes

Present

Ted Jibl, Matt Margolin, Jibari Simmons, Wendy Chisholm, Gregg Vanderheiden, Ben Caldwell, Dave MacDonald, Jason White, Michael Cooper, Roberto Scano, Matt May, Loretta Guarino Reid, Cynthia Shelly, Kerstin Goldsmith, Avi Arditti, Lee Roberts

Regrets

Charles Oppermann, Andi Snow-Weaver, Katie Haritos-Shea, Doyle Burnett, John Slatin, Gian Sampson-Wild

Action Items and Open Issues

• Action: michael, dave look at what is in best prctices and in the old Core Techniques doc and determine how similar they are. come up with criteria to decide which should be in best practices vs techniques.
• OPEN Issue: "current best practices" is a very used and loaded term. We should be sure to avoid or be careful how we use the terms.
• OPEN Issue: are best practices normative? This is a separate discussion re: conformance claims.
• OPEN Issue: not comfortable with only 2 sets, prefer 3
• Action: michael propose text of erratum to the list
• Action: kerstin and dave continue discussion of grammar on the list

Required success criteria vs best practise

Discussion of Potiential Criteria for placing items in success criteria vs. best practice - Gregg Vanderheiden, 29 May 2003

Issue: Reasons to have a set of criteria that is more specific

• if we have the criteria we can better address comments that we receive as well as decrease how much we argue amongst ourselves.
• we can better support decisions we have made.
• document how decisions are going to be made.
• if a decision is not made it accordance, then document it.

Resolved: delete the first item in the list and replace w/something more detailed about how we make decisions

Issue: "the best practice items might be testable or might not..." we should decide if they should or shouldn't be.

Discussion:

People seemed ok that best practices don't have to be reliably testable, as long as there is no conformance claim about anything that is not testable (i.e., only reliably testable items are normative). The idea was raised that perhaps we should have the minimum success criteria which are testable, a middle layer, and then best practice (which would not need to be reliably testable).

Other suggestions and questions:

• divide success criteria into human-testable and machine-testable
• how does this relate to gradient reporting?
• could move "non-minimum but not best practice" to parallel checkpoint in extended (idea was discussed and didn't seem to work, i.e., what if can't make it into checkpoint, have nothing beyond minimum?.

Resolved: stay with min/best practice unless we come across one that doesn't fit and then figure out what to do.

Issue: What is the difference between Required and Best Practice? Do best practices appear in guidelines/checkpoints or in techniques?

For example, how to write good alt-text is in Core Techniques. Currently, there areat least 16 sections in old core techniques. Writing good alt-text is similar to writing clearly and simply

In the Techniques Task Force, we have been working under the assumption that techniques support success criteria. We have not thought about the relationship between techniques and best practices. To write a technique there needs to be something in the guidelines. We assume we have to write at least one technique for each success criterion. Is this assumption correct? If so, do we also have to write a technique for each best practice? If not, then they could be more technique-like in character.

If the techniques guide implementation (because that is where developers go for information - rather than guidelines/checkpoints), then if a technique does not exist for a success criterion, it is likely the developer won't do anything in relation to that criterion.

Action: michael, dave look at what is in best prctices and in the old Core Techniques doc and determine how similar they are. come up with criteria to decide which should be in best practices vs techniques.

OPEN Issue: "current best practices" is a very used and loaded term. We should be sure to avoid or be careful how we use the terms.

OPEN Issue: are best practices normative? This is a separate discussion re: conformance claims.

OPEN Issue: not comfortable with only 2 sets, prefer 3

[#179] Meta "location" redirect

[#179] Meta "location" redirect from Michael Cooper, 28 May 2003

Do we have http techniques? No, but we're planning server-side techniques.

Resolved: remove from html techniques and move to server-side techniques

[B259] Proposal: remove default form text requirements

[B259] Proposal: remove default form text requirements - Michael Cooper, 27 May 2003

This checkpoint in WCAG 1.0 no longer seems to be relevant. It was created as a kludge for older user agents that are no longer commonly used.

Proposal 1: decide there will be no technique and guideline in wcag 2.0 that suggests putting placeholder text in forms. Publish a WCAG 1.0 errata that it is no longer applicable.

Proposal 2: write a technique that says, "this used to be a technique, no longer is. log issue."

Proposal 3: In Techniques docs, label some techniques as "deprecated."

Questions and comments:

• Will we recommend that people not do that?
• Is it negative for AT or UA?
• Are there ua issues w/presence of text in text input control?
• An older version of Jaws was broken and it was a problem. However, it was fixed in the next version.
• If we publish an erratum, it shows that WCAG 1.0 is still breathing.
• Need to support migration from 1.0 to 2.0. thus, might want to map 2.0 techniques to 1.0 (supports the "deprecated" techniques proposal)
• The "until user agents" checkpoints were mostly due to problems with AT that probably don't exist anymore.

Action michael: propose text of erratum to the list

[#248] Data and layout tables: identifying and marking

[#248] Data and layout tables: identifying and marking - Chris Ridpath, 29 May 2003

There are several proposals in this email, several issues were identified:

• Issue: Not sure that "add a statement that tables be used for entirely data or layout and not mixed" is feasible. Have seen outer table as layout and inner table is data.
• Issue: why a requirement that only one or the other on a page?
• Issue: summary only on first layout table?
• Issue: limiting # of characters in summary to x characters seems very english-specific. "keep succinct" depends on language and context.
• Issue: should only layout tables or only data tables have summary? for layout tables, summaries don't seem helpful.
• Issue: What about null summary fo layout tables? nothing happens differently, compared to images (and null alt-text)

Resolved: take back to Techniques Task Force

Active vs Passive Voice

Issue: Do we want to use passive or active voice? (the primary issue is to be consistent)

Long discussion about imperative, active, and passive.

Action: kerstin and dave continue discussion of grammar on the list

$Date: 2005/01/11 11:07:50 $ Wendy Chisholm