I've been doing the same wysiwyg-evaluation game and I immediately fell in love with WYMeditor when I saw it.
The biggest problem is that it doesn't yet work in Safari, but I think a lot of those you listed are in the same boat, because I've evaluated them all too.
I don't know -- it doesn't particularly excite me. All I care about is HTML; I don't need semantic markup beyond that. All most users care about is what they can see; what it means is implied, and only important in so far as the meaning is clear to the reader. Writers write for readers, not for computers.
That said, I am very interested in microformat support, which is kind of schema-related (but very loose schemas mixed in with content). I don't think anything has that now; part of why the community and code matters to us is because we want to build on this, we don't just want to use the product. A more semantic/abstract editor has some advantages here; but I'm not sure it's enough of an advantage, especially compared to a more mature product (and maturity really matters for this case, as there's tons of little usability issues).
One of the plugins I like for Xinha is the stylist plugin (to see it you have to look at the demo and select Stylist). I think there's some hidden potential in that, or something based on it.
Wow, I really like that WYSIWYM editor. That fits the web perfectly. I always thought that the problem with WYSIWYG editors is that the process is of "styling" the information, which doesn't really fit the way we deliver the web (CSS+XHTML). But the WYSIWYM definitely fits that model.
I'm sold.# Brantley Harris
It's actually somewhat surprising to me just how well most WYSIWYG editors respect CSS styles, with no special work to do it. I guess a testament to the underlying browsers' editing. Usually you have to pass a stylesheet in (so it can be loaded in the iframe that most editors use), but once you do that it's all just magic.
They still usually included non-semantic controls. But you can just take those controls out; again, adjusting the toolbar in most of these is easy. If you took those bad controls out (e.g., color, font), wouldn't you have something very much like wysiwym?
No, I don't think so. The idea of WYM is that you have something broken away from a context of the presentation. Instead you have something embedded in the structure of the very ideas. I think the practicality of it requires some formating elements to be displayed in the WYM-editor, i.e. bolding, italicizing, unordered lists, etc. But in the distant ideal those would just be highlighted and marked as such, rather than actually displayed as such.
In TinyMCE, at least, I can remove the controls from the toolbar, but someone could just paste something from word to circumvent my restrictions (especially annoying). Although, admittedly, I haven't looked for many fixes for that.
I think the idea of WYM could be expanded beyond just web-publishing. I imagine a collaborative screen-writing application. I should just go ahead and patent that right now, anyone have any capital for a startup?
WYM can't keep people from pasting HTML. It can try to clean that HTML (though it's rather tricky to do this, and given its age I suspect it doesn't). Several of the other editors have special code to try to clean up code pasted from Word; they also include different kind of cleaners, some which are applied on save and some which can be applied on paste.
However, I hate opinionated editors. Legacy HTML is important. HTML with differing opinions is important. If you are doing green-field content development I guess WYM could work, but I'm more interested in content development that builds on the content we already have. The words matter more than the form.