Re: Hidden file inputs [Was: Revert Request]

On Tue, Jan 31, 2012 at 11:43 PM, Charles Pritchard <> wrote:
> This ship has already sailed. As I understand it, Mozilla introduced it,
> WebKit is fixing bugs related to it.

Ah, that's a change of approach.

Relevant bug for others who are interested:

> It's a lot better than the previous situation of using: style="opacity: 0".

Easier for authors who want to style the file input, anyhow.

> Security is handled by the typical manner: restricting execution to the
> current gesture.
> That is, you can only send that click during a user gesture.

As long as gestures can be triggered by automation/accessibility APIs,
that sounds better than nothing.

>> As an aside, what's the advantage of including role="file button" in
>> your markup example?


>>    3. Constraining ARIA-Next to define the "file" role to mean "button
>> that causes a file dialog to open" rather than whatever the spec
>> editors think best at the time.
> That's the same issue as #2.


> I'm not constraining ARIA next at all, as I'm not putting "file" in a spec,
> I'm just sticking it in my current markup.

Compatibility with existing content, not specs, constrains spec
writers. You just need to spread your usage through enough of the web
that they are forced for web compatibility reasons to make their specs
match your usage. This has happened repeatedly in the evolution of web

> I don't believe "file" as a reference to the file system will become an ARIA
> role.
> I don't believe "folder" will either.

So you're prejudging what the "file" token should mean.

That's your prerogative, but what's the advantage of doing this?

Are you seeking standardisation of role="file"?

Benjamin Hawkes-Lewis

Received on Wednesday, 1 February 2012 00:25:33 UTC