Author: Dena S.

I compared Word Online, Word desktop, Google Docs, Gmail, Adobe Dreamweaver, the Confluence Wiki (used by wiki.illinois.edu), Drupal 7’s CKEditor implementation, Publish.illinois.edu’s TinyMCE implementation, and Box Note.

The analysis of the findings are below, followed by a table of data.

• Every editor surveyed files “normal” (or “body” or “paragraph”) with the headings.

• Every editor surveyed changes “headline” to its “normal” style when that is selected.

• Every editor surveyed doesn’t remove a link that’s in place when “normal” is applied over something with a link in it.

• These are behavior standards that people have reason to expect everywhere they see editor interfaces. That provides a strong established-behavior-based reason to follow these standards.

• These behavior standards don’t cause accessibility problems that we’re trying to actively change people’s behavior with.  So we don’t have correspondingly strong behavior-altering reasons to break these standards.

• We’re not trying to specifically block people from being able to create paragraphs the way we’re trying to specifically block people from being able to use the phrase “read more” in a link. “Read more” is getting active programming countermeasures to combat people’s link-making behavior. But we don’t have an accessibility problem caused by people using paragraphs to try to prevent.

• Providing interfaces that follow established expectations does provide an accessibility benefit for people with cognitive disabilities, in addition to a usability benefit for everyone.

• The majority of editors don’t remove “bold” formatting when “normal” is applied over something that is bold. 

• Word and Google Docs both do remove “bold” formatting.

• CKEditor’s default behavior mirrors the numeric majority of editors, so staying with its default behavior will reinforce the numeric majority’s definition of standard behavior.

(or)

• Word and Google Docs have a large amount of user mind share between them, and people likely spend more time in Word and Google Docs than the other editors (giving them a higher “time percentage” than “number percentage” of behavior-setting exposure time).  

• Different editors disagree in 4 different ways on how to handle it when a “bulleted” item is then assigned to “normal.”

• With no clear standard set, we wouldn’t be breaking any  clear expectations with whatever decision we make. 

• We can choose to leave CKeditor’s default behavior as the default here.

(or)

• If we’re looking to establish A11yFirst as something that’s clearly different than unmodified CKEditor, following Word’s behavior rather than unmodified CKEditor would be a way to emphasize the difference.