kInitialFrameDelay was added (by yours truly) to avoid having to "fix" a lot(?) of testcases when moving the SMIL engine over to "rAF-based" frame scheduling.
While it probably isn't something authors will notice (easily) it's a bit of a wart and a nuisance (to me at least...)
We should remove its use in SMILTimeContainer::Start and do something about the more newly grown usage in SMILTimeContainer::AdvanceFrameForTesting (open-coding probably if nothing else.)
Comment 1 by schenney@chromium.org
, May 25 2017