During drafting a monitoring methodology for compliance tests, training implementers of EN 301 549 and dealing with specific monitoring cases I haven't really understood intentions and rationale of so-called summarizing criteria (9.6, 11.5.2.3 and 11.8.1). Notably confusing is criterion 9.6 because despite detailed description that I have included into the Estonian methodology for compliance tests, implementers of EN 301 549 have frequently asked how to actually implement it.
From the viewpoint of monitoring based on the Commission Implementing Decision (EU) 2018/1524 these criteria will be non-compliant (i.e. results of procedure to check compliance with EN 301 549 will fail) anyway if minimally one criterion of a particular sample (precisely criteria 9.1 to 9.4 in case of criterion 9.6, criteria 11.5.2.5 to 11.5.2.17 in case of criterion 11.5.2.3 and criteria 11.8.2 to 11.8.5 in case of criterion 11.8.1).
There isn't need to repeat non-compliance because at least in Estonia we have to anyway report about every finding regarding non-compliances to participants in proceedings. I guess actually that this is partially the case in all EU members as well, because the 2018/1524 (precisely Annex II) guide to member states report regularly about
the findings regarding frequent or critical non-compliance with the requirements identified in the standards and technical specifications referred to in Article 6 of Directive (EU) 2016/2102.
Thus I am proposing to e-examine appropriateness of 9.6, 11.5.2.3 and 11.8.1 for websites (i.e. Table A1) and mobile applications (i.e. Table A2).
Designs
Child items
...
Show closed items
Linked items
0
Link issues together to show that they're related or that one is blocking others.
Learn more.
I certainly recognise the difficulties regarding testing these "summarizing" criteria and there is a high probability that one element of these complex might fail and therefore mean that the "summarizing" criteria cannot be fully met.
11.5.2.3 and 11.8.1 were unique requirements designed specifically for inclusion in EN 301 549. They do represent important contributions towards software accessibility, so removing them or significantly altering them is not something that should be done without some serious consideration.
Clause 9.6 also contains some important requirements related to website accessibility. Unlike the other two clauses, clause 9.6 is a direct copy of the WCAG 2.X Conformance Requirements. Prior to the publication of EN 301 549, many European countries and organisation required websites to conform to WCAG 2.0, which meant that the same challenges in testing and claiming conformance had to be faced before EN 301 549.
Initially EN 301 549 did not include these conformance requirements, but W3C were very keen to see clause 9.6 added to EN 301 549 so that conforming to EN 301 549 and to WCAG 2.0 were equivalent. Removing clause 9.6 would be likely to be very controversial!
Unfortunately I still don't understand, what will happen if we are going to remove 11.5.2.3 and 11.8.1 from the EN 301 549? As much as I understand, if ICT products and services have to comply with criteria of Annex A or another set of criteria that include criteria 11.5.2.5 to 11.5.2.17 and/or criteria 11.8.2 to 11.8.5, then they have to comply with 11.5.2.3 and/or 11.8.1 respectively. Thus, what is the uniqueness of 11.5.2.3 and 11.8.1?
9.6 is confusing even if we leave it as it is. Specific issue is with the requirement 9.1.2.4 (Captions (live)). 9.1.2.4 isn't listed in Table A.1, but on the contrary, 9.6 is listed in Table A.1.
A Web page that meets all of requirements 9.1 to 9.4, or where a Level AA conforming alternate version
(as defined in WCAG 2.1 [5]) is provided, will meet conformance requirement 1.
Thus if we have a rare situation, when a website of public sector body conforms with every requirement out of requirements 9.1 to 9.4 except 9.1.2.4, then how should an accessibility evaluator assess compliance with 9.6?
However I understand the rationale of 9.6 and it seems very reasonable. I don't know, if I simplify too much, but perhaps is a solution to rephrase requirement 9.6 and/or respective guideline for testing compliance (C.9.6) in line with the rationale (conforming to EN 301 549 and to WCAG are basically equivalent)?
For example, 1. step in procedure of C.9.6 may be: "Check that the web page conforms with every requirement out of requirements 9.1 to 9.4 (except 9.1.2.4, if the web page has to conform with Table A.1)." Moreover, the requirement needs perhaps a 4. note for implementers that explains how to actually implement it (as I wrote before, it's one of the frequently asked questions). I.e. the 4. note explains briefly that each conformance requirement is explained at WCAG Conformance section (analogous to wording of requirements that base on WCAG).
I understand your concerns about the "summarizing criteria" 11.5.2.3 and 11.8.1 being expressed requirements that are additional to the requirements that they are "summarizing". However, removing them would lose important explanations and qualifications that they contain. The points made in these two requirements should be retained. One possibility is to try to reword 11.5.2.3 and 11.8.1 to make them informative and not normative. The only other option would be to incorporate the important points raised in these clauses into each of the requirements that these are "summarizing".
The resolution of the concerns regarding clause 9.6 cannot be so easily addressed by the above proposal to modify the existing clause to make it non-normative. Clause 9.6 is a much more complex issue as it relates to:
The intention that the European Commission and the authors of the standard that compliance with clause 9 of EN 301 549 means exactly the same as compliance with WCAG 2.x and vice-versa.
Removing clause 9.6 would solve the first issue, but it would mean that the second issue can no longer be relied upon. It would mean that complying with EN 301 549 would be a less rigorous test than strict compliance with WCAG 2.x.
The decision to remove 9.6 is a policy choice that could only be taken if every party with a major interest in the accessibility of European websites, including the European Commission and W3C, are willing to accept that applying EN 301 549 will be less rigorous than applying WCAG 2.X.
Thank you again for the explanations. The idea to declare that compliance with clause 9 of EN 301 549 means exactly the same as compliance with WCAG 2.x is very sympathetic.
As I have gotten questions about the clause from those who are examining an Estonian methodology for testing conformance to EN 301 549 created by me, I rewrote a description of the clause in the methodology to make it more understandable. I have 3 findings from it:
I suggest adding a new note (currently it would be Note 4) to 9.6 that says what the intention of 9.6 is (to declare that compliance with clause 9 of EN 301 549 means exactly the same as compliance with WCAG 2.x). It would make the clause much more understandable (I described the intention in the description).
Despite the intention, we are still having an issue with WCAG SC 1.2.4, and the issue arises from the respective legislation: Like Note 4 on page 89 of the standard says, "Because the Web Accessibility Directive (EU) 2016/2102 "does not apply to live time-based media", the following requirements are not listed in Tables A.1 and A.2." (One of the following requirements is 9.1.2.4). Due to this, we have a contradiction: The 1st conformance requirement of WCAG (and clause 9.6) says "For Level AA conformance, the Web page satisfies all the Level A and Level AA Success Criteria", but it is impossible if we don't satisfy clause 9.1.2.4. Thus, it seems that it is correct to use a formulation like "WCAG 2.x level AA (except SC 1.2.4)" (I wrote to the description "criteria of WCAG that are transposed to clause 9 of EN 301 549 V3.2.1 (2021-03)").
I suggest adding a brief note to C.9.6 on how to reflect it in testing methodologies. For example, reflection of the 1st requirement ("One of the following levels of conformance is met in full.") is easy (at the end of the testing, an evaluator simply checks if all the applicable requirements of clause 9 are met), but how to test compliance with other requirements, i.e., I understand very well what the 2nd, 3rd, 4th, and 5th requirements mean, but I am not sure if I should do something more in addition to the tests that I already conducted.
Clauses 11.5.2.3 and 11.8.1
I support the proposal to make 11.5.2.3 and 11.8.1 informative and not normative. Thus, in my understanding, we should find a way to omit these from all annexes that are going to describe the relationship between EN 301 549 and the essential requirements of WAD and EAA (for example, the current Annex A).
As a member of the STFs that have written all versions of EN 301 549, let me try to provide additional explanations to these three requirements.
9.6 WCAG conformance requirements
Many people are not aware that meeting WCAG implies passing (i.e. "not failing") the success criteria under specific conformance conditions (the "conformance requirements"). These are additional requirements that developers/testers need to consider: that the full page meets all SC, that all steps in a process meet all the same SC, that there is no failure in "critical" success criteria ("non-interference"), and so on.
As the EN fully adopts WCAG, we must also include these conformance requirements. That is the purpose of 9.6. And so we must keep it (modified if necessary).
In addition, it does not make sense to modify any EN requirement due to legislation exemptions. The EN defines the requirements that an ICT has to meet to be accessible. And it is legislation that decides some circumstances in which some ICT does not need to be accessible (like "live media"). The notes explaining applicability of legislation should be in the specific annexes (currently annex A), not in the main part of the EN.
11.5.2.3 Use of accessibility services
This clause is a single requirement: "applications must use the platform accessibility services". It is not informative.
It was written with a reference to the more detailed requirements 11.5.2.5 to 11.5.2.17 to clarify what needs to be done if the app runs in an operating system that does not provide such services. But if the operating system is providing the necessary accessibility services, then the requirement is simple: "use the services".
We might want to refine the writing, but I don't think that it has to become non-normative.
11.8.1 Content technology
This is the only requirement that in my opinion could be transformed into non-normative text, because it only explains that conformance to 11.8.2 to 11.8.5 depends on the possibilities of the output format used by the authoring tool.
But if 11.8.1 becomes non-normative, then we need to update the preconditions of 11.8.2 to 11.8.5 to add something like "Where the ICT is an authoring tool and the output format enables...".
I understand the point that meeting WCAG implies passing the success criteria under specific conformance conditions. As I said, I suggest adding a brief informative note to C.9.6 on how to reflect it in testing methodologies. I.e., how the evaluator makes sure in the testing process that the 4th requirement ("Only accessibility-supported ways of using technologies are relied upon to satisfy the success criteria.") is compliant.
The note about the applicability of legislation (i.e., the issue with WCAG SC 1.2.4). I am not sure if Annex A is the best place for that note. At the moment, notes about Annex A and its tables are on pages 89–90, but the insertion of 9.6 in Table A.1 is on page 93 (it is very easy to not notice the respective note). Perhaps it is an appropriate solution to add the note next to C.9.6 on page 148, since at the end of the day, this is the issue of assessment. The first step of the procedure is to check that the Web page satisfies all the Level A and Level AA Success Criteria of WCAG. Clause 9.1.2.4 doesn't have to be satisfied if the Web page has to follow the requirements of WAD and EAA (both of them should be aligned as explained in Issue #75 ). Thus, the first check fails, and the entire result of C.9.6 will fail. If we have the note there, we won't have the issue.
The note about the intention of 9.6. It seems appropriate to add it to the main part of the EN since it is only informative and makes it much more understandable (as you said: "Many people are not aware that meeting WCAG implies passing (i.e. "not failing") the success criteria under specific conformance conditions (the "conformance requirements"). ").
11.5.2.3 Use of accessibility services
Refining 11.5.2.3 and/or C.11.5.2.3 seems necessary, since at the moment it gives the impression that the main purpose of it is to summarise requirements 11.5.2.5 to 11.5.2.17, not to state that "applications must use the platform accessibility services".
9.6 and WAD exclusion of live media
Thank you @puskar@martineznorm@pluke for dealing with the challenge of 9.1.2.4 (regarding live time-based media) being excluded from the WAD and yet required by clause 9.6 which is listed as relevant to WAD in annex A.
I support the idea of adding informative note(s) about this. Unsure about the best place for this (9.6, C.9.6 or annex A). I believe a note will be found by most people if it is placed in 9.6, but probably this isn't a big issue in practice, so the important thing is to have the documentation somewhere. Those who really search for it should be able to find it even if it's in C.9.6 or annex A. Especially if there are online sources (such as this GitLab thread) that can be found using search engines, and that can direct people to the note.
9.6 and conforming alternate versions
Another aspect of 9.6 that I believe needs further clarification in EN301549 is the formulation "where a Level AA conforming alternate version (as defined in WCAG 2.1 [5]) is provided".
If the EU intention, as stated previously in this thread, is "that compliance with clause 9 of EN 301 549 means exactly the same as compliance with WCAG 2.x and vice-versa", then clause 9 should also allow for conforming alternate versions. Despite the inclusion of 9.6 I do not believe this is the case right now: A website may pass C.9.6 using a conforming alternate version, but it will not pass whatever other clauses there are under section 9 that caused the need for a conforming alternate version.
The EN301549 Conformance chapter (clause 14) contains a hint that the combination of two ICTs might "together meet more requirements of the standard when one item complements the functionality of the other". But there are no clear preconditions - as in WCAG chapter 5 - for how this could be realised. Nor is clause 14 referenced in the mapping tables for WAD.
Do you agree that this is a disalignment between WCAG and the EN301549, and that it needs to be changed?
Comments from respondents in questionnaire to members in the Wadex group:
We see this as duplicating other requirements that refer to WCAG 2.1. Because if any requirement from WCAG 2.1 fails, this requirement must also be violated automatically. Should a penalty be imposed twice for the same violation?
WCAG conformance requirements are very relevant and contain additional requirements in comparison with WCAG chapters 1-4 and add an option to meet the requirements by having a conforming alternative version.
This requirement should definately not be dropped but to be added to apply to non-web software and non-web documents too.
The W3C WCAG2ICT Task Force spent a significant amount of effort trying to construct an equivalent version of "conforming alternative version" for software but were unable to do so. The primary difficulty is that the conforming alteratvie version would typically be a different web page, but due to the way that W3C identifies complete applications as the equivalent of a single web page, it is difficult to understand what a conforming alternative version could possibly be (a completely different application!).
@rodriguezasc and @puskar will attempt to propose solutions for different issues raised here about clauses 9.6, C.9.6, 11.8.1, and C.11.8.1 (issues about 11.5.2.3 and C.11.5.2.3 that were raised here got resolved as noted in issue #239 (closed)).
Regarding 9.6 and C.9.6, they have detected 5 different topics that are necessary to address:
"Summarising" effect of the clause and understandability of the intention of the clause
Ambiguity of C.9.6 in testing methodologies
Clause 9.6 and the WAD exclusion of live media
Clause 9.6 and conforming alternate versions
Application of WCAG conformance requirements to non-web documents and non-web software that provides a user interface
Regarding 11.8.1 and C.11.8.1, the most viable approach at the moment seems to be transforming 11.8.1 into non-normative text.