W3C home > Mailing lists > Public > public-html@w3.org > June 2010

Re: CfC: Adopt ISSUE-105 canvas-usemap Change Proposal to add usemap attribute to the canvas element

From: Charles McCathieNevile <chaals@opera.com>
Date: Wed, 23 Jun 2010 10:00:31 -0700
To: "Sam Ruby" <rubys@intertwingly.net>, "Ian Hickson" <ian@hixie.ch>
Cc: "HTML WG" <public-html@w3.org>
Message-ID: <op.veriu5w3wxe0ny@widsith.local>
On Wed, 23 Jun 2010 01:10:06 -0700, Ian Hickson <ian@hixie.ch> wrote:

> On Tue, Jun 15, 2010 at 4:05 AM, Sam Ruby <rubys@intertwingly.net> wrote:
>> We have a change proposal to modify 4.8.10 the canvas element section  
>> of the HTML5 specification to allow the usemap  attribute to be applied
>> to the canvas element:
>> http://www.w3.org/html/wg/wiki/ChangeProposals/addimagemaptocanvas
> Sorry for the delay, I missed this call. I object to this proposal,
> ... and it
> has a vastly higher implementation cost than is justified by the
> authoring benefit derived.

What is the implementation cost? We do this in Opera right now. A smoke  
test shows that FF and Safari do as well. Those are the three browsers I  
have to hand, but I would be interested in the evidence that this is an  
implementation cost (rather than having to change existing browser code to  
change the way the Web works today).

> Furthermore, because this would get used so rarely, browser vendors
> would likely fail to prioritise this when fixing bugs, and we'd end up
> with this feature being perennially buggy.

I don't understand the evidence or logic behind the claim. It seems to  
work now, it matches what browsers already do with object (FF and Opera)  
and images which are rendered onto the screen. So I don't understand  
either where the bugs might come from, or why it would be untested.

It is clearly far simpler to write an image map for a few regions of a  
canvas than to manage hit-collection

> It should also be noted that the change proposal does not give enough
> detail for the spec to be edited to address the proposal. Merely
> making the "usemap" attribute legal on <canvas> doesn't do anything to
> make usemap="" actually work on <canvas>.

The single sentence about usemap which apparently is good enough for the  
object element is probably about good enough for the canvas element as  
well. I would have thought that was obvious, but if you can't understand  
what is intended I agree that the change proposal should give you  
sufficiently explicit editorial guidance. Note that my change proposal for  
ISSUE-74 includes a resolution to ISSUE-105 that is effectively adopting  
Steve's change proposal.

(That proposal also includes changes to the image map specification, which  
appears to be based on something that is a nice idea but represents a  
clear implementation cost for at least Opera, Safari and Firefox which do  
not implement the new HTML5 image maps in general. I believe the issue is  
orthognal to whether image maps work on canvas or not, but I wonder if  
that belief is correct).



Charles McCathieNevile  Opera Software, Standards Group
     je parle français -- hablo español -- jeg lærer norsk
http://my.opera.com/chaals       Try Opera: http://www.opera.com
Received on Wednesday, 23 June 2010 17:01:17 UTC

This archive was generated by hypermail 2.4.0 : Saturday, 9 October 2021 18:45:20 UTC