- From: Randall Leeds via GitHub <sysbot+gh@w3.org>
- Date: Mon, 23 May 2016 22:36:44 +0000
- To: public-annotation@w3.org
Yes, just followed up to same the same about `Selection` as you. I am not sure I favor assuming that the `Selection` API is used to stringify a selection. In Hypothesis, for instance, I implemented the text quote selector by inspecting the text values of the ranges retrieved from `getSelection().getRangeAt(n)` for `0 <= n < getSelection().rangeCount`. In the past, the Selection API was used, but it made anchoring more compute intensive. The easiest way to guarantee the same normalization was to traverse the entire DOM and select every node. The alternative used now can just call `document.body.textContent` and let the browser return all the text content efficiently. I suppose both approaches seem equally valid to me, I just wonder if we really gain much by specifying additional normalization. -- GitHub Notification of comment by tilgovi Please view or discuss this issue at https://github.com/w3c/web-annotation/issues/221#issuecomment-221116720 using your GitHub account
Received on Monday, 23 May 2016 22:36:45 UTC