- From: Marcus Geelnard <mage@opera.com>
- Date: Wed, 01 Aug 2012 06:19:31 +0000
- To: Tony Ross <tross@microsoft.com>
- Cc: Chris Wilson <cwilso@google.com>, Raymond Toy <rtoy@google.com>, Jussi Kalliokoski <jussi.kalliokoski@gmail.com>, "public-audio@w3.org" <public-audio@w3.org>
Citerar Tony Ross <tross@microsoft.com>: >> From: Marcus Geelnard [mailto:mage@opera.com] >> >> I think that splitting the spec into two layers would make it easier to >> focus on what's really important to get right in the spec (i.e. the >> core), and getting that part ironed out and testable ASAP. That way we >> can hopefully lift the "experimental" tag from at least the most >> important parts of the API, while continuing to stabilize the next level. > > I definitely think there's value in identifying which features we > think are most important for developers so we can ensure these > stabilize sooner. I'm not advocating where that line should be > (maybe it's even most of the current spec), but we certainly should > have some form of consensus on it, if for nothing other than to aid > in discussions around potential additions to the standard. Yes, I think the key point here is to identify and prioritize the work on the "core". Splitting the spec into levels would just be a very distinct way of doing that. > To start, I do think the flexibility provided by JavaScriptAudioNode > makes it an important part of that core feature set, regardless of > what other nodes/capabilities we decide to include. Also anywhere we > do draw a line we should ensure that the presence or absence of the > next level of support can be easily identified via feature > detection to reduce the confusion/burden placed on developers. Agreed. I personally think that the "architecture" part (context, graph routing, automation, etc) should be in the core feature set, plus a minimal set of nodes that make it possible to implement the majority of sound processing nodes in a straight forward manner (though possibly with less performance than with specialized nodes). That way we would be quite sure that the API is future proof and easily extensible. /Marcus
Received on Wednesday, 1 August 2012 06:20:07 UTC