w3c / strategy

team-strat, on GitHub, working in public. Current state: DRAFT
151 stars 45 forks source link

Privacy Field Attributes - a W3C Member solution (new Startup Member) #254

Open JAlanBird opened 3 years ago

JAlanBird commented 3 years ago

Peter Shikli from Access2Online would like to have the following considered and an extension to HTML.

They have a group providing telephone support to the DMV. They're looking at next-generation digital options for that, with formalized chat and screen sharing the two best contenders. The screen sharing will be faced with the privacy concerns that is the subject of this offer to help form the necessary standards. Without the W3C's input, they will have to respond with their best designs. This will include use of a privacy factor attached to sensitive fields as described below.

With screen sharing becoming popular, particularly as driven by the virus-driven growth in teleworking, we need a standard by which private form field values are to be protected. One use case is a customer seeking screen sharing support filling in an ecommerce order form that displays a credit card number. Another is an insurance form with private medical fields. We need a standard methodology by which such fields can be designated as private and their values replaced by asterisks much as is done with passwords. A candidate approach is to designate such fields with a privacy attribute, perhaps an integer from 0 to 9, such that browsers can recognize the privacy attribute and display according to website publisher defaults and/or user settings as well as use cases such as screen sharing.

nother point worth making is how we developers have started solving the problem without a W3C standard. We've implemented a form with a "Privacy" toggle button. When the user clicks that, the values of designated fields are replaced by asterisks. The button turns into "Public" and clicking it returns the previous value displays. One advantage is that this also insures privacy in over-the-shoulder public settings.

Among the disadvantages are that there is no standard by which this communicates with screen share software or anything else. A confused user, a common type seeking support, may forget to press the Privacy button and then the screen share will display private info. Also, each developer will implement this privacy feature in their own way, some good, some bad, but resulting in the industry disorder the W3C was to prevent.

samuelweiler commented 3 years ago

My understanding is that screen-sharing tools look at display buffers with no particular awareness for the semantics of the content the are showing - the architecture does not give them a way to even understand that they're seeing a text field, nevermind make any sense of (invisible) content labels.

It sounds like they're looking for a change in the architecture of screen sharing?

If I have misunderstood, a proof of concept ("running code") would go a long way to demonstrate the utility of this approach and make the case for standardizing the labeling.

JAlanBird commented 3 years ago

I'd like to understand from the Strategy Team who would be available to see this or do we carve 15 minutes from a Strategy call and ask Accessi2Online to demo it for the whole team. I plan to agenda+ this for next week's Strategy call to have the next steps discussion.

dontcallmedom commented 3 years ago

there are lots of ongoing discussions in the WebRTC Working Group around screen sharing and some of its associated semantics - https://github.com/w3c/mediacapture-screen-share/issues