Added "Duplicated text" as guidance to fulfil the 1.4.5 Images of Text success criterion in the Understanding Document

[giacomo-petri]

Closes: #3755

This is a proposal to include "duplicated text" as a sufficient technique to satisfy the 1.4.5 Success Criterion, addressing users needs to personalize text presentation while also ensuring sustainability for companies (as entirely replacing text images with real text often results in inaction).

Updates:

• Included a section in the understanding document to consider this approach.
• Provided an example.
• Introduced a sufficient technique.

[patrickhlauke]

@stevefaulkner asked about clarification if the technique implies the visible alternative text needs to always be visible, or if would be good to also show an example where the text is in a disclosure widget or similar. would be good to provide a separate example with this

[dbjorge]

I don't think I'm on board with this. I agree that this likely should be an exception to 1.4.5, but I don't think it is an exception today; I think if we agree that this should be an exception, we should be exploring an erratum, not adding a technique that tries to define an exception where none exists.

[patrickhlauke]

Additional text is an accessible alternative, in effect. Don't think it becomes an exception per se, more an example

[patrickhlauke]

i.e. this is a case of

alternatives to part of a page's content are considered part of the page when the alternatives can be obtained directly from the page, e.g., a long description or an alternative presentation of a video.

https://www.w3.org/TR/WCAG22/#cc2

so wonder if that even needs a technique as such, OR if the technique should clearly explain that it's not trying to make a new exception, but that it demonstrates the point there of the "Full pages" note 1

[stevefaulkner]

putting links here for info purposes:

https://www.w3.org/TR/WCAG22/#images-of-text

This would be an appropriate technique if the image of text is considered essential 

For example infographics. WebAIM has an example which has a structured text alternative below the graphic

[giacomo-petri]

Following our previous discussions regarding lengthy PR requests, if further discussion arises here, we can move the discussion back to #3755.
Since I garnered some consensus in the initial issue and received no additional feedback, I submitted a proposed solution as a PR. However, I am open to revisiting the original issue for further refinement.
Nonetheless, I still believe that clarification is required on this matter.

[patrickhlauke]

To put it succinctly, after considering this further, this to me feels less like a technique to pass 1.4.5, but an example of "in isolation, that image with text fails 1.4.5, but because there's an accessible alternative, overall the page/sample passes". Which would feel odd to have as a technique, even though I can see how the info is useful to have. Maybe, instead of a technique, worth adding a new example/note to the 1.4.5 understanding itself, which makes the scenario clear ("sometimes, there's just no way around having images with text, but then you should provide an accessible alternative such as a visible text rendition of the same information")

[giacomo-petri]

@patrickhlauke,

Do you believe the current proposal could work by merely eliminating the technique? If so, happy to simply remove the technique.

[patrickhlauke]

@giacomo-petri I think even just the change to understanding may need a subtle re-orienting ... maybe it's semantics, maybe it's just me, but i think it's not so much that

Meeting this success criterion is still deemed acceptable by duplicating the meaning of the image of text through textual format

but perhaps more along the lines of "if it's not possible to meet this SC directly, you must provide an accessible alternative - such as duplicating the meaning/content of the image in text..." or something (maybe even linking to https://www.w3.org/TR/WCAG22/#cc2)

[bruce-usab]

...allows content creators to incorporate both an image and a caption. While an image of text

I think "while" should be "when" here. Also is automated or policy?

[bruce-usab]

I would like a rendered version please.

[giacomo-petri]

Hi @bruce-usab,

As decided by the group, I'll first remove the technique and adjust the SC wording accordingly. I'll tweak the note slightly and add an example instead of a technique.

[giacomo-petri]

@patrickhlauke, is it better now?

[patrickhlauke]

@giacomo-petri yup i think in principle i'm happy with these changes now :)

[bruce-usab]

I prefer Alastair's suggestion.

Phrasing "it could add or altering" reads oddly to me.

[giacomo-petri]

Updated with agreed version and fixing some typo.

[alastc]

From the meeting: The additional examples do not actually pass the criterion, the page is passing by an alternative.
It would be useful to separate those examples into a separate list and highlight that they don't pass.

E.g.

The following examples pass the criterion:
[current list]
The following examples do not pass the criterion, but the page would pass because there is an alternative.
[new list]

[mbgower]

It is unclear what you are trying to link to with this URL: https://www.w3.org/TR/WCAG22/#cc2

There is no definition of accessible alternative. Are you making the case that the page can conform although it fails 1.4.5, by meeting 1.1.1?

[giacomo-petri]

It is unclear what you are trying to link to with this URL: https://www.w3.org/TR/WCAG22/#cc2

I'm referring to NOTE 1, which states:

For the purpose of determining conformance, alternatives to part of a page's content are considered part of the page when the alternatives can be obtained directly from the page, e.g., a long description or an alternative presentation of a video.

---

Are you making the case that the page can conform although it fails 1.4.5, by meeting 1.1.1?

Not really. SC 1.4.5 and 1.1.1 address different needs. SC 1.1.1 is primarily intended for people who can rely on text alternatives that don't need to be visible, while SC 1.4.5 requires an option to visually customize the text. Therefore, if the same text is presented close to the image, it addresses the requirement.

[mbgower]

IMO, the wording needs to be clearer in these additions.
I believe there was a agreement that text that appears on the same page (not just an ALT) as the image of text (i.e., a caption that states the same information) can be considered to pass 1.4.5, since all information is presented as text.

It's less clear to me that folks would agree that an exact but undisplayed ALT for an image of text passes.

[giacomo-petri]

@mbgower,

The intent is indeed to require a visible text alternative, not an invisible one like an alt attribute.

Does adding "visible" to the sentence (emphasis on """visible"""):

If it is not possible to replace an image of text with actual text or to allow users to customize the image, the author must provide an accessible alternative by duplicating the meaning or content of the image in """visible""" text form.

solve the problem?

[giacomo-petri]

Note: new examples seem already covering this bit clearly.

[mraccess77]

As SC 1.4.5 is aimed at sighted users - alt text which is not displayed would not pass in my opinion. If it would, then there would be no material point in SC 1.4.5 as 1.1.1 already requires a text equivalent for images of text.

[patrickhlauke]

randomly dropping here: as the SC is scoped to situations where images of text are used rather than text, then the case where images of text are used in addition to text, then the SC doesn't apply anyway since it's not the rather than

[giacomo-petri]

Based on yesterday's discussion, I have drafted a new proposal with a different approach to avoid any ambiguity or conflict with the normative portion of the SC:

Proposal:

Suggested change

	<p>If it is not possible to replace an image of text with actual text or to allow users to customize the image, the author must provide an [accessible alternative](https://www.w3.org/TR/WCAG22/#cc2) by duplicating the meaning or content of the image in text form.
	The alternative could replicate the text verbatim. If converting the visual image to text results in a loss of meaning or readability, the author could add or alter some words to convey the style, emphasis, or any other meaning that was lost.
	</p>
	<p>An image of text accompanied by visible text that replicates its content verbatim is exempt from this success criterion, as the image is not used to convey information but serves a decorative purpose, with the text providing the actual information.
	Moreover, the text does not always need to replicate the image of text content exactly. Verbatim replication might result in a loss of meaning or readability. Authors can modify or add words to convey style, emphasis, or other meanings that could be lost in a direct replication.
	</p>

If the group is ok with this proposal, I will also revise the examples to better align with the changes.

[patrickhlauke]

I would actually suggest dropping this, and the previous change (lines 47-50), and maybe instead just adding something like the following (which I think is getting to the same point as we discussed):

<div class="note">
  <p>The Success Criterion is scoped to situations where <a>images of text</a> are used <em>rather than</em> text. If a page has an image of text, but also contains visible text that duplicates the textual content of the image and conveys the same meaning, then this Success Criterion does not apply.</p>
</div>

[mbgower]

I'm not sure that it doesn't apply, so much as that it meets the requirement.
If text is used, it's met. I'm wondering if we're able to align on some concept and wording like this:

Note: The Success Criterion is intended to address situations where images of text are used rather than text. Where images of text are used in addition to text to convey the same information -- where both are presented to the user -- this success criterion is met. This allows authors to convey content using any styling they desire, while also presenting the information in text, which can then be manipulated by users to make it more distinguishable.

[mbgower]

I agree that I'm inclined not to include new content that would open the Pandora's box regarding how much the text may deviate from the same information provided in the image of text before this SC is no longer met via this rationale.
If we decide to strip out some of the other additions, I'm thinking this might work best as a note as a third-to-last paragraph in the Intent section.

[mbgower]

I'm having difficulty picturing this. We could mabye include an example? Otherwise, I'm inclined to not add this last example.

[patrickhlauke]

Following more discussions in last week's WCAG 2.x backlog meeting, I filed a clean counter-proposal #4021

[alastc]

Suggest we wait for Michael and Giacomo to get back before resolving this one.

[alastc]

Superceded by #4021