Thread by @FrankElavsky

Exhausting. While this might be a cause for celebration for some, my preliminary audit shows that this is yet again another tool for easily creating inaccessible data experiences.

No SR or keyboard access, no semantic control of marks or their relationship to one another, etc

https://twitter.com/mbostock/status/1389611128339144704

As just naked DOM stuff, this means that yet again the onus is not on the creator of the library but the consumer to do accessibility. Why do we continue to make it easy to make inaccessible things?

Disappointed because this solves technical barriers for some, but produces many.

Why do big names/groups/companies in this space continue to innovate exclusionary tools, libraries, and resources?

These fast and easy solutions create more accessibility problems than they solve. We are long overdue for accessibility and inclusion in the wide field of data.

EG We are 22 years past the pub of Grammar of Graphics. What a rich potential to expand on the theory + produce a true semiology of what the marks mean.

But instead we have spend two decades creating technical implementations, leaving many behind in a growing technological gap.

We can specify commands to a computer in a myriad of ways in order to make geometry do stuff. This keeps getting easier.

But we still have a significant gap to bridge describing the thing that is PRODUCED by these geometric relationships. In many ways, this work is now harder.

If Plot can infer how to create a chart with limited input, can we not imagine futures that meaningfully infer descriptions, alt-text, aria-labels, keyboard navigation, and high-level relationships based on what is produced?

(Sounds like good PhD material...)

Don't get me wrong, Observable Plot is neat. It will rapidly speed up the analytical workflows of able-bodied people.

But what about people with motor impairments (13-15%), uncorrect-ed/able vision impairment (21-28%), cognitive disability/impairment (9-11%), or the blind (4%)?

Nearly one quarter of all people are estimated to have some form of disability or another that affects their life daily.

Yet a full 100% of us will experience disability, impairment, limitation, or exclusion at some point in our lives due to temporary or situational context.

To really clarify that last comment:
EVEN YOU (yes YOU reader) will experience disability, limitation, or exclusion at some point in your life.

Disability inclusion and accessibility is personal to YOU.

Inclusive tools help everyone do their job, including you.

I am simply too tired of new technological workflows that don't consider disability. I can't celebrate Observable Plot like so many of you will.

Using Observable Plot not only limits who you can share your work with but also who can participate in this work.

Good alternatives that at least consider accessibility:
Highcharts (excellent accessibility options, low-level)

Visa Chart Components (great accessibility but high level, IE declarative chart types)

Vega-lite (okay accessibility, easy to use in Observable, low-level)

Sources on disability and inclusion:
Microsoft's Inclusive Design 101:
download.microsoft.com/download/b/0/d…

WHO Executive Summary for the "World Report on Vision:" apps.who.int/iris/bitstream…

CDC on "Disability Affects All of Us:" cdc.gov/ncbddd/disabil…

Post-script: I expect some will encounter this thread and find it unfair ("Plot is meant to be concise," "Bostock et al aren't experts in a11y," "since it is DOM-level a11y isn't impossible," etc).

Simply put: the work that must be done was not. These excuses don't matter.

Post-post-script: I also expect the vast majority of the community will simply be apathetic to accessibility and simply will not care.

This (perhaps) is the real reason it is so exhausting to exist in this deeply ableist space.

If we don't take action now, litigation will.