- From: Doug Schepers <schepers@w3.org>
- Date: Sat, 28 Aug 2010 15:08:23 -0400
- To: Anne van Kesteren <annevk@opera.com>
- CC: www-dom@w3.org
Hi, Anne- Anne van Kesteren wrote (on 8/28/10 1:55 PM): > On Sat, 28 Aug 2010 19:48:22 +0200, Doug Schepers <schepers@w3.org> wrote: >> Anne van Kesteren wrote (on 8/13/10 10:14 AM): >>> http://www.w3.org/TR/DOM-Level-3-Events/#event-type-scroll -- in my >>> limited testing for specifying this event as part of CSSOM View I found >>> that it simply uses the Event interface (not UIEvent) >> >> It is a UI Event, so we believed it would be better placed there; the >> implementation difference it trivial, and we didn't think moving it >> would negatively affect legacy content. If you have evidence >> otherwise, and the group agrees that's better, I could move it back to >> the basic Event interface. > > What is the advantage? As .detail means nothing and .view is useless. The advantage I see is a more logical and consistent framework on which to build future events, and easier for developers to learn. In several of the event interfaces, there are properties that only have meaningful values for a subset of the events on that interface, but which are still grouped logically, so this has precedent. From a developer and script library perspective, there may be reasons to treat UIEvents differently, and this makes it easy to detect that category of events (rather than reply upon a hardcoded list of event types). > Just seems like additional complexity for no reason. Isn't any additional complexity is negligible in terms of implementation and performance? However, if it emerges during LC that the group feels this should be categorized as part of the basic Event interface, I will change it. >>> and does bubble >>> when dispatched on the Document object (to Window, and bubbles is true). >> >> Is this a special case for the scroll event on the Document object? >> Could you share your tests with us? > > It works differently for the Document object, yes. (I don't have > anything I can share in test form at this point.) Okay, thanks for the heads-up; I didn't realize this before, and it's probably the reason 'scroll' was listed as bubbling before. I've noted this in the spec, in the event type definition and the event table: http://dev.w3.org/2006/webapi/DOM-Level-3-Events/html/DOM3-Events.html#event-type-scroll Regards- -Doug Schepers W3C Team Contact, SVG and WebApps WGs
Received on Saturday, 28 August 2010 19:08:26 UTC