[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Public WebGL] Some WebGL draft feedback
Tim Johansson wrote:
On 2010-01-13 20:43, Kenneth Russell wrote:
The only implementation that would render correctly without
synchronization is one where all graphics API's are implemented in
software running on a single processor together with the application,
i.e. no parallelism possible. This is also the only kind of
implementation where everything would not slow down when mixing APIs.
Such an implementation is extremely unlikely.
I agree that incorrect rendering is not good as it is not well
specified. If an implementation manages to get it looking right
without calling the sync function that behavior will quickly become
required since many authors will forget to call the sync functions.
Incorrect rendering. The exception checks would have the same
performance and tentacle-spreading issues I raised before.
I don't think incorrect rendering results are an acceptable
alternative. I also dispute whether the check to see if the given
rendering context is "current" upon each graphics call would have any
perceptible performance impact. We currently have and need such
checks in the WebGL implementations for Safari and Chrome and they
are negligible in cost.
That means the check is required and I don't see any real benefit of
requiring a sync function to be called instead of just syncing
implicitly. IMO we should not add functions to make authors think
about everything that might be slow in some implementations.
org:HI Corporation;Graphics Lab, Research & Development
adr:Higashiyama 1-4-4, Meguro-ku;;Meguro Higashiyama Bldg 5F;Tokyo;;153-0043;Japan
tel;work:+81 3 3710 9367 x228
tel;fax:+81 3 5773 8660