Intent to Deprecate and Remove: navigator.vendor and navigator.vendorSub

[Simon Pieters]

*Primary eng (and PM) emails*
sim...@opera.com

*Summary*
Remove navigator.vendor and navigator.vendorSub from Blink.

*Motivation*
navigator.vendor is sometimes used for browser sniffing and is thus
causing some compatibility pain. navigator.vendorSub just returns the
empty string and isn't being used by anyone. These are not specified in
the HTML spec.

*Usage information from UseCounter*
I haven't instrumented.

I did however check http://webdevdata.org/ data-dec2012.zip (doesn't
include external scripts) where I basically found three instances (out of
top 10,000 sites' front pages) checking for "Google" as a way to sniff for
Chrome.

*Compatibility Risk*
A few pages might break as they will fail to sniff for Chrome. Opera has
considered this to be low enough risk to change the value from "Google,
Inc." to "Opera Software ASA", and we then considered the idea of dropping
it altogether. Evidently Google found it low enough risk to change from
"Apple Computer, Inc." to "Google, Inc.", although that has caused a few
bugs:

https://code.google.com/p/chromium/issues/detail?id=10498
https://code.google.com/p/chromium/issues/detail?id=58659
https://code.google.com/p/chromium/issues/detail?id=21059

In the long run, it seems like it would result in less compatibility
problems if navigator.vendor was dropped.

*Row on feature dashboard?*
No.

*Requesting approval to remove too?*
Yes.

--
Simon Pieters
Opera Software

[Eric Seidel]

I believe we need usage data. I expect these may be very commonly
used by JS libraries.

[Adam Barth]

I would very much like to remove these, but I agree with Eric that the place to start is with UseCounter data.

Adam

[Ola P. Kleiven]

UDM menu system [1] breaks in Opera 15+ as it sniffs for
navigator.vendor=='Google Inc.' Don't know how much that is used these
days, but development seems halted.

1) http://www.udm4.com

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



--
Ola P. Kleiven
WebStandards
Opera Software

[Simon Pieters]

On Tue, 20 Aug 2013 01:14:02 +0200, Adam Barth <aba...@chromium.org> wrote:

> On Mon, Aug 19, 2013 at 3:39 PM, Eric Seidel <ese...@chromium.org>
> wrote:
>
>> I believe we need usage data. I expect these may be very commonly
>> used by JS libraries.
>

> I would very much like to remove these, but I agree with Eric that the
> place to start is with UseCounter data.

Yes, you are right. I've asked a colleague to set up a UseCounter.

cheers

[Christian Biesinger]

Do other browsers implement it? (IE, FF)

-christian

[Mike Sherov]

Firefox populates empty strings, IE does not have them.

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

--

Mike Sherov
Chief Technologist
SNAP Interactive, Inc. | Ticker: STVI

http://snap-interactive.com | http://ayi.com

[Daniel Bratell]

Den 2013-08-20 17:57:01 skrev Mike Sherov <mike....@gmail.com>:

Firefox populates empty strings, IE does not have them.

I've filed http://code.google.com/p/chromium/issues/detail?id=276813 for eventual removal (and further discussions and patches).

/Daniel

[Eric Seidel]

I still believe test-first, then remove is the safer pattern.
However, given this information, I'm less concerned about usage of
this webkit-only feature and would be ok with a "remove first, see who
breaks" approach here. Up to you.

[Paul Irish]

Github code search reveals a good amount of projects using navigator.vendor. 

https://github.com/search?p=1&q=navigator.vendor&source=cc&type=Code (the total result count seems high, but it includes many duplicated projects)

Many small things in there, nothing I would have a problem breaking.

It looks like Keynote's HTML Export uses it to identify Safari and serve it an upgraded slide-viewing experience. So that will probably be the source of many (and potentially most) hits with UseCounter.