People start hacking around the limitations of the specification in order to keep supporting it. Sillyness. Especially when you consider that there is browser support for far easier non-hacky solutions. Take Opening new windows with JavaScript, version 1.1 for example, or Flash Satay: Embedding Flash While Supporting Standards. How about just using the target attrubte and the embed element? Now I’m not 100% sure about the use cases for target="_blank", but I guess there are some for applications at least.
The embed element is cool but only if you don't want content fallback.
I understand your way of thought, Anne, but I can´t say I agree. And I´ve been feeling that a discussion like would rise for quite some time now, but I would not guess that you´d be the provocoteur.
My view on the target-issue is simple - HTML is trying to be a monkey while it´s just a tortoise.
Browser behaviour is way out of HTML´s field of expertise, this is Java-/VBScript´s assignment.
About embed I´m not sure, but I´m generally sceptical to the common use of Flash on the Internet...
It's an issue of fundamentalism. Where do your priorities lie: 100% semantic/standards submission at the cost of productivity (replacing embed with a javascript that outputs basically the same thing) or productivity at the cost of some bizarre pursuit of pat-yourself-on-the-back validation? As long as you're one or the other, the opposing side's priorities are tough to consider.
Anne I encourage you to read Including Multimedia Object.
There are two problems with embed, the one which has been already cited in the example, and there is another which is due to patents. It was the reason at the start to not use this specific name.
Interesting thread, with the same old debate, about Embed element