Re: [editing-explainer] replaceText input event could span multiple Typing Nodes (#50)

None of this is breaking existing functionality, as contentEditable=typing is a completely new concept and everything existing will live on in contentEditable=true. We have made no promises that this will allow for everything that has been possible with other editing environments in the past. Some editors currently implement spellchecking through third part services instead. There are even entire solution(s?) written in javascript to handle spell checking purely in javascript.

As for the last point: editors are complex and browser makers *cannot* know everything that is going on in them. But given that no direct access to spell checks is granted to JS, it is reasonable to suspect that things within the same text node are letters in the same word if there is no space between the,. Beyond that, it is not reasonable to make any guesses on the part of the browser what is going on. For example. Take the following:

```
<p>Is this a wo<span class="a">rd</span>?</p>
```

The browser would probably guess that "word" is a single word. But what if the CSS stated something like:

```

span.a {
display: block;
border: 5px solid green;
position: absolute;
top: 0px;
}
```

Suddenly it's something entirely different. Would the browser know? Maybe, maybe not. From the perspective of the JS, it would most likely seem out of control. You would need A LOT of specification to explain when something crossing two text nodes is still a word. And then Javascript programmers would go through the implementations and find bugs that help them turn cross-element word detection on and off.

So therefore: I don't think it is a good idea. But if you still think it is a good idea, could you write a spec to determine when words crossing element boundaries are still words and edit this spec referring to the other. Then we can vote on the two proposals at the F2F meeting. 

---
Reply to this email directly or view it on GitHub:
https://github.com/w3c/editing-explainer/issues/50#issuecomment-105289630