- From: Michael Gower <michael.gower@ca.ibm.com>
- Date: Wed, 4 May 2022 17:52:23 +0000
- To: Alastair Campbell <acampbell@nomensa.com>, "Bradley-Montgomery, Rachael" <rmontgomery@loc.gov>, "WCAG list (w3c-wai-gl@w3.org)" <w3c-wai-gl@w3.org>
- Message-ID: <SA0PR15MB4032B353322FA11752AE4CA5DEC39@SA0PR15MB4032.namprd15.prod.outlook.com>
Having dug into this a bit, I wonder if one option would be to use the COCO categories. They had a list of 272 candidate categories and selected 91 https://scholarphi.semanticscholar.org/?file=https://arxiv.org/pdf/1405.0312v3.pdf https://tech.amikelive.com/node-718/what-object-categories-labels-are-in-coco-dataset/ There are obvious drawbacks to being so prescriptive. On the other hand, given CAPTCHA is likely to draw from this list, I think it is a supportable and meaningful starting point. Has anyone investigated using this, and are there any results of that exploration? Mike From: Michael Gower <michael.gower@ca.ibm.com> Date: Wednesday, May 4, 2022 at 9:43 AM To: Alastair Campbell <acampbell@nomensa.com>, Bradley-Montgomery, Rachael <rmontgomery@loc.gov>, WCAG list (w3c-wai-gl@w3.org) <w3c-wai-gl@w3.org> Subject: Re: Accessible Authentication SCs, WCAG 2.2 It helps with context, but I don’t think it resolves the reality of the SC wording. There is no official response at the end of issue 1902, which is normal for us with normative text issues. It has a response from COGA TF about 1/6 of the way through the conversation<https://github.com/w3c/wcag/issues/1902#issuecomment-867697712>. Given its location in the conversation, it represents a point from which the SC evolved. The final comment from you says using a dictionary definition is “mentioned above”, but that is the only time in the entire issue where “dictionary” appears, and it’s not mentioned in the Understanding document. Issue 1256 is useful to understand how we arrived at where we did with the exception for recognizing common objects in images (from the research John R cited: “Picture-based CAPTCHA tests do not pose many of the problems faced by users with learning disabilities”). That would actually be good to explain (or at least referenced) in the Understanding document. 1256 also has some discussions to do with what Jon A just raised (and to a degree addresses those, IMO). A quick summary of that in the Understanding document would be helpful. Ultimately, I still think we need to define what constitutes common objects. Mike From: Alastair Campbell <acampbell@nomensa.com> Date: Wednesday, May 4, 2022 at 8:28 AM To: Michael Gower <michael.gower@ca.ibm.com>, Bradley-Montgomery, Rachael <rmontgomery@loc.gov>, WCAG list (w3c-wai-gl@w3.org) <w3c-wai-gl@w3.org> Subject: [EXTERNAL] Re: Accessible Authentication SCs, WCAG 2.2 Hi Michael, We discussed and agreed that back in September: https://www.w3.org/2021/09/21-ag-minutes#item02 We also had an issue specifically on that topic: https://github.com/w3c/wcag/issues/1902 ˛ ˛ ˛ ˛ ˛ ˛ ˛ ˛ ˛ ˛ ˛ ˛ ˛ ˛ ˛ ˛ ˛ ˛ ˛ ˛ ˛ ˛ ˛ ˛ ˛ ˛ ˛ ˛ ˛ ˛ ˛ ˛ ˛ ˛ ˛ ˛ ˛ ˛ ˛ ˛ ˛ ˛ ˛ ZjQcmQRYFpfptBannerStart This Message Is From an External Sender This message came from outside your organization. ZjQcmQRYFpfptBannerEnd Hi Michael, We discussed and agreed that back in September: https://www.w3.org/2021/09/21-ag-minutes#item02<https://www.w3.org/2021/09/21-ag-minutes#item02> We also had an issue specifically on that topic: https://github.com/w3c/wcag/issues/1902<https://github.com/w3c/wcag/issues/1902> Which was closed by the update agreed in the meeting. Another useful background discussion was: https://github.com/w3c/wcag/issues/1256<https://github.com/w3c/wcag/issues/1256> A couple of clips: “[The SC] doesn't explicitly define common objects, but as mentioned above, we're just using the dictionary definitions of those terms.” “Crap captchas (e.g. the car vs van) affect everyone, although I'm sure it's [more difficult] if you have certain disabilities.” Does that help? -Alastair From: Michael Gower I’m concerned we have no definition of “common objects”. It seems problematic as part of the normative text, and I believe it must be defined or addressed in some way. Is a list appropriate/feasible? The Understanding document discusses some of the considerations for common object. Based on how the definition of common object is tackled, more context likely needs to be added there. Additional concerns with this SC (that can be resolved at a later point) are: * A better explanation of the authentication process, as inferred by the SC, to better orient users to the common considerations (i.e., many people would not think of password recovery as part of an authentication process) * A better focus on what the author responsibility is – we need clear guidance on what is required of someone creating a page. Right now, ya really have to dig for it. Once that’s established, then the doc can discuss how that author piece integrates with ATs, user agents, etc. * the Understanding document in a few places is to some degree a list of techniques (which are not then crafted as separate techniques). It would be better to create additional sufficient and failure techniques (which will contribute to the prior bullet) Mike From: Bradley-Montgomery, Rachael <rmontgomery@loc.gov> Date: Tuesday, May 3, 2022 at 5:58 PM To: Alastair Campbell <acampbell@nomensa.com>, WCAG list (w3c-wai-gl@w3.org) <w3c-wai-gl@w3.org> Subject: [EXTERNAL] Re: CFC - Accessible Authentication SCs, WCAG 2.2 +1 From: Alastair Campbell <acampbell@nomensa.com> Date: Tuesday, May 3, 2022 at 7:37 PM To: "WCAG list (w3c-wai-gl@w3.org)" <w3c-wai-gl@w3.org> Subject: CFC - Accessible Authentication SCs, WCAG 2.2 ˛ ˛ ˛ ˛ ˛ ˛ ˛ ˛ ˛ ˛ ˛ ˛ ˛ ˛ ZjQcmQRYFpfptBannerStart This Message Is From an External Sender This message came from outside your organization. ZjQcmQRYFpfptBannerEnd +1 From: Alastair Campbell <acampbell@nomensa.com> Date: Tuesday, May 3, 2022 at 7:37 PM To: "WCAG list (w3c-wai-gl@w3.org)" <w3c-wai-gl@w3.org> Subject: CFC - Accessible Authentication SCs, WCAG 2.2 Resent-From: <w3c-wai-gl@w3.org> Resent-Date: Tuesday, May 3, 2022 at 7:33 PM Call For Consensus — ends Monday May 9th at midday Boston time. The Working Group has previously discussed the WCAG 2.2 SC Accessible Authentication, and the AAA version with no exception, and they need to be approved by CFC. Both can be previewed in the editor’s draft: https://w3c.github.io/wcag/guidelines/22/#accessible-authentication<https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fw3c.github.io%2Fwcag%2Fguidelines%2F22%2F%23accessible-authentication&data=05%7C01%7Cacampbell%40nomensa.com%7Cb48b623b4e4c49075b1808da2dd7b71a%7Cebea4ad6fbbf43bd8449c56e26692c35%7C0%7C0%7C637872701767083240%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=nJu0Xj33TQMb5ZHriqUUmtUSQdWdGaKS5LGGxF3nIBo%3D&reserved=0> They were last discussed Dec 3rd: https://www.w3.org/2021/12/03-ag-minutes#t08<https://www.w3.org/2021/12/03-ag-minutes#t08> The change history is here: https://github.com/w3c/wcag/commits/main/guidelines/sc/22/accessible-authentication.html<https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fw3c%2Fwcag%2Fcommits%2Fmain%2Fguidelines%2Fsc%2F22%2Faccessible-authentication.html&data=05%7C01%7Cacampbell%40nomensa.com%7Cb48b623b4e4c49075b1808da2dd7b71a%7Cebea4ad6fbbf43bd8449c56e26692c35%7C0%7C0%7C637872701767083240%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=RX8dXDMyHNIZAUo6yiVQuA8cwP4muLIrHSA3VnHiGcQ%3D&reserved=0> The 17 issues opened and dealt with are listed in the survey: https://www.w3.org/2002/09/wbs/35422/wcag22-accesssible-auth-updates/results<https://www.w3.org/2002/09/wbs/35422/wcag22-accesssible-auth-updates/results> If you have concerns about this proposed consensus position that have not been discussed already and feel that those concerns result in you “not being able to live with” this decision, please let the group know before the CfC deadline. Kind regards, -Alastair -- @alastc / www.nomensa.com<http://www.nomensa.com>
Received on Wednesday, 4 May 2022 17:52:51 UTC