- From: Jason White <jason@jasonjgw.net>
- Date: Thu, 15 May 2014 20:24:40 +1000
- To: public-indie-ui@w3.org
James Craig <jcraig@apple.com> wrote: > Substantial edits made today/tonight covering the @uittrigger idea we've been discussing over the past few months. > > In addition to the rest of the edits, I'd appreciate a critical, technical review of this new section, entitled "UI Triggers." Thank you, James, for your excellent work in specifying the new feature. Consider the following comments "post-heartbeat" material - that is, they needn't be addressed before publishing the heartbeat draft. 1. The text states explicitly that UIManipulationRequestEvent types are not included in the set of tokens available as values of uitrigger, because additional parameters not expressible as simple tokens are needed in these cases. It follows that the motivating example for the uitrigger feature, the slider (operated by move events targeting the slider thumb) isn't covered. It needs to be addressed, however, then reinstated as an example in the spec. As I remember, the trigger element responds to move requests by dispatching value change requests, and we need to define the relationship between deltas in the move request and the the corresponding value change. 2. The text, as drafted, states that the trigger element need not be focusable if, for example, there is an access key assigned to the operation. This makes sense if focus means "keyboard focus", but I'm wondering whether there are any other input methods that rely on focus (e.g., sequential navigation in a touch interface) rather than merely on point-of-regard. On the systems that I have used, sequentially navigating UI controls in a touch interface simply changes point-of-regard, not focus, so the text as it stands may be fine - we're only concerned about focus in relation to keyboard input, but I would still suggest checking for possible counter-examples where this could be a problem. I don't think speech input relies on focus either, for instance, thus perhaps all non-keyboard inputs are covered without the element's having to be focusable. 3. There are minor errors worthy of cleaning up before publishing the new section. This point probably should be addressed before the heartbeat draft. 4. I think informative text should be added that explains the differences between uitrigger and uiactions, and gives the rationale for each. I expect these attributes to be a source of confusion unless their respective purposes are clearly delineated. As mentioned, I don't want to stand in the way of publishing a heartbeat draft in the agreed time-frame.
Received on Thursday, 15 May 2014 10:25:10 UTC