- From: Bristow, Alan <Alan.Bristow@elections.ca>
- Date: Thu, 16 Nov 2023 12:59:47 +0000
- To: Mark Magennis <Mark.Magennis@skillsoft.com>, w3c WAI List <w3c-wai-ig@w3.org>
- Message-ID: <YQBPR01MB10383D397EC3506212383383CFCB0A@YQBPR01MB10383.CANPRD01.PROD.OUTLOOK.CO>
Hi Mark, Thank you for your reply. This context and detail, added to other feedback, sounds increasingly like the problem; some bad HTML and NVDA making a guess at where to recommence and getting it wrong. Thank you so much for that context and detail, with the other feedback, it really helps solidify my approach now. I will try hard to remember to reply to this thread if an "Ah ha!" moment comes and I find a certain culprit; in case it's helpful to anyone finding the archive (https://lists.w3.org/Archives/Public/w3c-wai-ig/2023OctDec/0066.html). Thanks again! Regards, Alan . . . . - . . - - - Alan Bristow ( he / him / il ) Web Developer / Développeur Web Elections Canada / Élections Canada alan.bristow@elections.ca<mailto:alan.bristow@elections.ca> ________________________________ From: Mark Magennis <Mark.Magennis@skillsoft.com> Sent: 16 November 2023 07:11 To: w3c WAI List <w3c-wai-ig@w3.org> Subject: Re: Screen reader operation appears strange on multi-page form Ce message a été envoyé par un expéditeur externe. Veuillez faire preuve de prudence et ne pas cliquer sur les liens ou ouvrir les pièces jointes à moins de reconnaître l'expéditeur et de savoir que le contenu est sûr. This message was sent from an external sender. Please exercise caution and do not click links or open attachments unless you recognize the sender and know the content is safe. Hi Alan, In the past three years I've logged three separate bugs against my company's software for NVDA starting reading from the footer in a new page. The product is an SPA. In all three cases, focus was at the top of the page but NVDA began reading from the start of the footer. Two of those bugs were resolved by some general work we did on organising focus on page load throughout the application. The other was fixed separately by a developer who is no longer here so I can't ask him what he did to resolve it. But just to point out that this may not be exclusive to your situation but may be something that NVDA does under a specific set of circumstances. As I'm sure you know, screen readers often use heuristics to decide what to read or do when the code is invalid or unexpected and I wonder if this may be happening here. NVDA may be assuming, due to some set of circumstances, that you should be in the footer. Kind of like when you remove the focused element from the page. The browser shifts focus to the <body> but 'bookmarks' the place it was as the 'tabbing resumption point', so when you Tab it goes to the element following the one that was removed, but if you arrow using a screen reader it starts from the top of the page. I could imagine a screen reader, if it was aware that this had happened, deciding to start browsing from the tab resumption point rather than the focus point, thinking that this is probably the best user experience. I'm not saying this is what is happening but it may be some similar attempt by NVDA to put you where it thinks you ought to be rather than where you are. Sometimes screen readers' heuristics turn out to be bad guesses and they are usually made because there is something incorrect, incomplete, or ambiguous in the code. Mark ________________________________ From: Bristow, Alan <Alan.Bristow@elections.ca> Sent: Wednesday 15 November 2023 17:16 To: w3c-wai-ig@w3.org <w3c-wai-ig@w3.org> Subject: [EXTERNAL] Screen reader operation appears strange on multi-page form If anyone with screen reader (SR) experience knows the answer to the following I would be most grateful - perhaps this is a misuse of this list, if so my apologies and please ignore. Scenario A survey is split over 5 HTML pages. When page 1 has been completed NVDA announces "Next" as I land on a BUTTON element that submits page 1. Activating the button, page 2 loads, and NVDA commences announcing the page content. However, the content it reads appears to consist of the contents of the TITLE tag, followed by what came after the similar spot where it left off on page 1, the content of the FOOTER element. This misses out all the new content that page 2 presented. Tried with JS turned off. Tried in incognito with no extensions having access. Tried with the mouse pointer nowhere near the page (I usually navigate by keyboard (arrows / tab) when testing with SRs). I wondered if perhaps each of the five FORM tags, one in each page, needed unique IDs, or something similar, but my research has failed to touch on this initial guess. Any comments gratefully received. Regards, Alan . . . . - . . - - - Alan Bristow ( he / him / il ) Web Developer / Développeur Web Elections Canada / Élections Canada alan.bristow@elections.ca<mailto:alan.bristow@elections.ca>
Received on Thursday, 16 November 2023 12:59:57 UTC