RE: Error correction and "steps"

I have 3 questions regarding this proposed criterion:

If the user can go back one screen at a time to review the steps is that acceptable? My read of this is that going back a step will meet and this and in conjunction with redundant entry will allow for re-entry of data if going back were to not save something.

Does this address the situation where the user can go back and forth in steps until they reach a final screen that has a cancel or submit button?  That is does there need to be an edit or back button on the final submit step?     Previously they had the ability to change but at some point in the process you can no longer edit.  I've seen situations like this on airline sites.

Can restarting the process by canceling or closing out if data is not lost be a way to meet this?  For example, I do this while shopping on Amazon - if I get to the last step and realize I made a mistake I can just close out at the very last step and then go back to my cart and checkout again with my items still there although I may need to make a selection for delivery again if it's different than default.  I'm asking these questions because many sites don't have an edit or go back on the final step although up until this step you can do that.

Jonathan

From: Alastair Campbell <acampbell@nomensa.com>
Sent: Thursday, June 11, 2020 7:29 AM
To: WCAG list (w3c-wai-gl@w3.org) <w3c-wai-gl@w3.org>
Subject: Error correction and "steps"

CAUTION: This email originated from outside of the organization. Do not click links or open attachments unless you recognize the sender and know the content is safe.

Hi Everyone,

I was putting one of the final WCAG 2.2 SCs into github, and realised that having been through that big conversation on 'steps' for redundant entry, we have a whole other SC which also uses it.

In fact, it had a definition of steps that we all apparently agreed! (Towards the end of one of the long meetings in March.) Unfortunately, I had completely forgotten about it.

The discussion on Redundant Entry (without 'steps') is on Tuesday, and I think it's worth bearing Error correction in mind then.

The current PR for Error Correction has removed the term 'steps' from the normative text and most of the understanding doc:
https://raw.githack.com/w3c/wcag/wcag22-error-correction/understanding/22/error-correction-processes.html

Wilco had a concern about things like chat windows, but I think of each chat entry as a "submission" which you are previewing before you send it. If you consider the chat as a process, or part of a process then the chat part would come under "unless the information cannot be modified for logical... reasons"
Changing your previous chat entry would not be logical once the other person has received it.

Apart from that, I'm wondering whether people think that removing steps helps (because we really struggle to define that), or whether it makes the SC wider?

Kind regards,

-Alastair

--

www.nomensa.com<http://www.nomensa.com/> / @alastc