@RHo could you please link your Figma mock(s) for this when you have a moment?

After discussion with @MMiller_WMF, we've decided to pause adding this alt text field as an additional (albeit optional) input for users inside of the Add image task until we see higher completion rates (and lower revert rates) in the Add image task.

@KStoller-WMF @RHo could we consider this for the Section-Level-Image-Suggestions project? I'm concerned about further rollout of a feature that prevents accessibility best practices, and has a negative impact on the accessibility of content that newcomer editors produce.

I'm open to considering it Q3/Q4 when we focus on image work.

Note the proposed initial design below is for the proposal to enable entry of an additional alt text field as *optional* as part of the "Add image" structured task (in the caption step).
Another avenue to address this issue may be for another structured task entirely for adding ALT text to any image missing this field.

Since we already know newcomers are struggling with the caption step, it might be worth considering that second option... but I suppose we don't need to decide on implementation now.

In T296820#8351371, @KStoller-WMF wrote:

I'm open to considering it Q3/Q4 when we focus on image work.

Note the proposed initial design below is for the proposal to enable entry of an additional alt text field as *optional* as part of the "Add image" structured task (in the caption step).
Another avenue to address this issue may be for another structured task entirely for adding ALT text to any image missing this field.

Since we already know newcomers are struggling with the caption step, it might be worth considering that second option... but I suppose we don't need to decide on implementation now.

Given newcomers are already struggling with caption, perhaps we should consider the alternative proposed in the task which is to separate out a different structured task entirely instead of, or in addition to adding a second caption-like ALT text field for people to enter.

In T296820#8352636, @RHo wrote:

In T296820#8351371, @KStoller-WMF wrote:

I'm open to considering it Q3/Q4 when we focus on image work.

Note the proposed initial design below is for the proposal to enable entry of an additional alt text field as *optional* as part of the "Add image" structured task (in the caption step).
Another avenue to address this issue may be for another structured task entirely for adding ALT text to any image missing this field.

Since we already know newcomers are struggling with the caption step, it might be worth considering that second option... but I suppose we don't need to decide on implementation now.

Given newcomers are already struggling with caption, perhaps we should consider the alternative proposed in the task which is to separate out a different structured task entirely instead of, or in addition to adding a second caption-like ALT text field for people to enter.

That sounds interesting, but seems outside the scope of our current year and I suspect it wouldn't get done next year either, given the amount of work needed to implement a new structured task.

Some alternative lighter weight implementations:

• a button (perhaps placed on the image) that allows users to add alt text, if they want
• Setting the alt field with the contents of the caption field
• implementing some default text, like "refer to caption" in the alt field

Those would all be relatively straightforward to build; the first one is measurable, if we want to see how users intentionally engage with it.

In T296820#8359200, @kostajh wrote:

In T296820#8352636, @RHo wrote:

In T296820#8351371, @KStoller-WMF wrote:

I'm open to considering it Q3/Q4 when we focus on image work.

Note the proposed initial design below is for the proposal to enable entry of an additional alt text field as *optional* as part of the "Add image" structured task (in the caption step).
Another avenue to address this issue may be for another structured task entirely for adding ALT text to any image missing this field.

Since we already know newcomers are struggling with the caption step, it might be worth considering that second option... but I suppose we don't need to decide on implementation now.

Given newcomers are already struggling with caption, perhaps we should consider the alternative proposed in the task which is to separate out a different structured task entirely instead of, or in addition to adding a second caption-like ALT text field for people to enter.

That sounds interesting, but seems outside the scope of our current year and I suspect it wouldn't get done next year either, given the amount of work needed to implement a new structured task.

Some alternative lighter weight implementations:

• a button (perhaps placed on the image) that allows users to add alt text, if they want
• Setting the alt field with the contents of the caption field
• implementing some default text, like "refer to caption" in the alt field

Those would all be relatively straightforward to build; the first one is measurable, if we want to see how users intentionally engage with it.

I like the idea of having the optional ALT text provide an ability to "use caption text", and/or to tone down the CTA for the ALT text further as an initial test. One thought and one question:

• We may want to test whether adding optional ALT text field could actually help improve captions, because people might be able to better understand what a caption should include when articulated as "describe how this image relates to the article" vs ALT text as "Describe what's inside the image for a non-sighted person".
• What is the current patroller attitude to ALT captions? Is it something that will be more, less, or equally scrutinised/reverted?

Regarding patrolling: default and community-built patrolling tools are using diffs, that are wikitext based, so they can monitor the alt text. Visual diffs cover the case as well. If caption and alt text are the same, most people will delete the alt text.

@KStoller-WMF I'm moving this to Triaged as we haven't made any progress in moving the task forward. It is listed under T293711: [EPIC] "Add an image" Iteration 2 so hopefully we can return to it.