- From: Joe D Williams <joedwil@earthlink.net>
- Date: Sun, 31 May 2009 13:36:33 -0700
- To: "Ian Hickson" <ian@hixie.ch>, "Larry Masinter" <masinter@adobe.com>
- Cc: "HTML WG" <public-html@w3.org>
from Subject: was RE: HTML interpreter vs. HTML user agent Sunday, May 31, 2009 10:58 AM Ian Hickson wrote: > ...though that's not been requested by implementors so far. Hi, speaking as an author-implementor, that is an author attempting to implement content using an embedded scriptable object plugin player, all I ask is that the W3C host browser makes it so that the user of the html5 tool is presented with an experience I can have some hope of predicting. Then I can plan for what will happen if the content is presented just exactly as I would wish, and to plan for however many levels of fallback seem feasible depending upon the user preferences and the user interface possiblities. Over the years, the <object> element has become the element of choice, due to the fact that more W3C browsers are really trying to accomplish a meaningful design that allows free and secure live communication between the DOM host and the active embedded object. VRML/X3D players have mostly agreed, or are agreeing, on the kind of host behaviors and services they would like to see and there is a mostly agreed upon sets of preferred DOM techniques and 'standard' <param> pairs used to control the plugin runtime. Yet the X3D players are also wavering on some simple decisions, such as a consensus on a <param> to pass the target content url. As a result, the current main web3d.org embedding example duplicates the data attr and the src param. That is where a user just walking up to this would be confused. Why do that? Of course since I have mostly followed how the <object> element and associated <param> element behavior has evolved, especially over the last 7 or so years, with new and improved W3C UAs, more 'open' vm possibilities that want to be an <object> emerging, and the fantastic improvements in W3C browser technology and robustness, I know it is better than ever! Overall, an author depending upon the <object> element for interactive multiple media has a better chance of success than at any time ever. It will only get better with deeper integration of interactive hyperlinked multiple media as the WWW advances technically. Of course I need to look more at and catch up on some surrounding <object> issues like sandboxing and seamlessing and sniffing, but starting out, here is one simple suggestion: * Obsolete using <embed> as a fallback for <object>. Move <embed> to Obsolete, like <applet>. There are no more processors that do not recognize <object> but recognize <embed> Sure, <embed> came first but it became archaic and is fully replaced and extended by <object>. Fairlly easy transform to replace <embed> with <object> or provide meaningfull fallback. Focus on just one of these, the advanced case, <object> will improve secure support for plugin handling and web content authoring tools and templates. Thanks to All and Best Regards, Joe http://www.web3d.org/x3d/wiki/index.php/X3D_and_HTML5
Received on Sunday, 31 May 2009 20:37:16 UTC