Re: ChannelMergerNode mixing semantics seem weird in dynamic situations

On Wed, Jan 16, 2013 at 6:21 PM, Robert O'Callahan <robert@ocallahan.org>wrote:

> ChannelMergerNode seems a bit fragile if the number of channels on an
> input is changing dynamically. The way it's defined, if the number of
> channels on input 0 changes dynamically, then the channels of inputs > 0
> move up or down in the channel layout. The same thing would happen if a
> node is connected or disconnected to input 0 while other nodes are
> connected to inputs > 0. In fact, the spec doesn't say what happens if
> input 0 has nothing connected but inputs > 0 do.
>

It's true that the behavior is tricky if the input(s) are changing
dynamically, so the developer would have to be careful to account for it.
 But, I believe it's well-defined.  I think in the vast majority of useful
cases, the inputs would be static in their channels and thus would present
a stable output.

To answer your question about what happens if input 0 has nothing connected
to it, it just doesn't influence the output at all and is skipped.


>
> It seems ChannelMergerNode would also interact poorly with the
> auto-removal of nodes from the graph (which is still itself a big problem
> ... I posted about that a while ago).
>
> I think it would be much simpler and more reliable, and easier to
> implement, if the number of channels taken from each input by
> ChannelMergerNode was constant for the lifetime of the node. For example,
> we could pass this as an additional optional parameter to
> createChannelMerger, defaulting to 1.
>

This node is one of the simplest we have in WebKit, and the implementation
is about 1.5 pages of code.  I'm happy to share some of the approaches
we've taken.


>
> Rob
> --
> Jesus called them together and said, “You know that the rulers of the
> Gentiles lord it over them, and their high officials exercise authority
> over them. Not so with you. Instead, whoever wants to become great among
> you must be your servant, and whoever wants to be first must be your
> slave — just as the Son of Man did not come to be served, but to serve,
> and to give his life as a ransom for many.” [Matthew 20:25-28]
>