The building analogy is no longer suitable if you're suggesting the ramp suddenly fails to accommodate new wheelchair designs. The wheelchair design should be your new focus of scrutiny. If the new wheelchair fails to conform to EXISTING ramps found everywhere, the wheelchair is the problem, not the ramp.
Ramps are built according to a strict building code. Your suggestion the ramp might need modifying in only the Twitter building by a dedicated team, doesn't add up when considering all ramps in all buildings.
> there isn’t a point where you’re done
My point is accessibility is inherently part of the design process anyway. Designers are not just making things pretty, clueless to accessibility. After all, Twitter is primarily text. You can always recruit outsiders from time to time, who will stress test and audit. That's a good way to keep on top of accessibility, by getting actual users who need accessibility features, to evaluate your service. But full time in-house? No.
> After all, Twitter is primarily text. You can always recruit outsiders from time to time, who will stress test and audit. That's a good way to keep on top of accessibility, by getting actual users who need accessibility features, to evaluate your service. But full time in-house?
This is like saying you don’t need ML specialists because it’s just math, or that we don’t need those expensive security people because Accenture told the CIO they could us experts any time we need them.
Accessibility is far more than just “can a screen reader produce text for this element” — starting from the fact that there are many accessibility concerns to which the issue of media forms is irrelevant but going much deeper into how you organize that information. Even if we ignore the prevalence of non-text content, Twitter isn’t a single chunk of prose but the much harder problem of a lot of complex elements which update frequently. You can’t treat this as some quarterly exercise where a consultant runs a scanner - there’s a lot of behavior which must be designed in and tested continuously.
When you talk about meeting outside users, yes, that’s good. Guess which team organizes that? Guess which team makes sure you get representative users? Guess who helps make sure developers and designers get the training and time they need?
Ramps are built according to a strict building code. Your suggestion the ramp might need modifying in only the Twitter building by a dedicated team, doesn't add up when considering all ramps in all buildings.
> there isn’t a point where you’re done
My point is accessibility is inherently part of the design process anyway. Designers are not just making things pretty, clueless to accessibility. After all, Twitter is primarily text. You can always recruit outsiders from time to time, who will stress test and audit. That's a good way to keep on top of accessibility, by getting actual users who need accessibility features, to evaluate your service. But full time in-house? No.