Intent to Deprecate and Remove: precision="float" attribute on <input type="range">

Philip Jägenstedt

unread,

May 27, 2014, 5:33:15 AM5/27/14

to blink-dev

Primary eng (and PM) emails

Summary

Remove the precision="float" attribute on <input type="range">, which is non-standard and can be replaced by step="any".

Motivation

10 years ago, this attribute was added to WebKit: http://trac.webkit.org/changeset/6311

I haven't been able to find it any spec, if it was it ought to have been here: http://www.w3.org/Submission/web-forms2/

5 years ago, support for step="any" was added to WebKit: http://trac.webkit.org/changeset/51159

These have the same effect, so the non-standard precision="float" can be removed.

Compatibility Risk

In this ad-hoc test, only WebKit and Blink appear to support the attribute:

http://software.hixie.ch/utilities/js/live-dom-viewer/saved/3041

Combined with the fact that usage is low, risk seems rather low.

Alternative implementation suggestion for web developers

Use step="any".

Usage information from UseCounter

http://www.chromestatus.com/metrics/feature/timeline/popularity/46

Usage appears to be slowly declining.

Entry on chromestatus.com and/or MDN

None.

Requesting approval to remove too?

Yes. The deprecation period doesn't seem useful, given the already low usage and lack of support in non-WebKit browsers.

PhistucK

unread,

May 27, 2014, 6:03:24 AM5/27/14

to Philip Jägenstedt, blink-dev

This either seems to have been mistakenly checked in, or it just was not mentioned in the commit message for some reason... funny.

☆PhistucK

To unsubscribe from this group and stop receiving emails from it, send an email to blink-dev+...@chromium.org.

Eric Seidel

unread,

May 27, 2014, 1:40:04 PM5/27/14

to PhistucK, Philip Jägenstedt, blink-dev

lgtm

TAMURA, Kent

unread,

May 27, 2014, 7:24:17 PM5/27/14

to Eric Seidel, PhistucK, Philip Jägenstedt, blink-dev

LGTM2.

I prefer having deprecation period for one or two releases because this attribute has been available for 10 years.

--
TAMURA Kent
Software Engineer, Google

Ojan Vafai

unread,

May 27, 2014, 8:27:46 PM5/27/14

to TAMURA, Kent, Eric Seidel, PhistucK, Philip Jägenstedt, blink-dev

LGTM3

PhistucK

unread,

May 28, 2014, 1:30:44 AM5/28/14

to TAMURA, Kent, Eric Seidel, Philip Jägenstedt, blink-dev

That makes sense.

☆PhistucK

Philip Jägenstedt

unread,

May 28, 2014, 8:38:23 AM5/28/14

to TAMURA, Kent, Eric Seidel, PhistucK, blink-dev

Sorry, I saw "LGTM2" and "LGTM3" on my phone and sent out the prepared
CL, which has now landed:
https://codereview.chromium.org/301563002/

Should I revert and deprecate instead?

Philip

Eric Seidel

unread,

May 28, 2014, 9:40:21 AM5/28/14

to Philip Jägenstedt, TAMURA, Kent, PhistucK, blink-dev

I doubt it matters, especially given the way it was landed (hidden in
another commit) and how no other browser supports it. WebKit-only
sites which depend on this will break by their <input type='range'>
being integers instead of floats and will have a trivial fix of using
an explicit step value (like step='any') instead (and then magically
working in all other browsers).

If someone could find documentation on the internet mentioning
precision=float, that could be convincing evidence that we should keep
it for a while. I wasn't able to find any mention of precision=float
outside of this thread.

http://stackoverflow.com/questions/19011861/is-there-a-float-input-type-in-html5
http://blog.isotoma.com/2012/03/html5-input-typenumber-and-decimalsfloats-in-chrome/

TAMURA, Kent

unread,

May 28, 2014, 6:49:02 PM5/28/14

to Philip Jägenstedt, Eric Seidel, PhistucK, blink-dev

I don't think you need to revert it. Sudden removal is also acceptable.

Reply all

Reply to author

Forward