Gervase Markham wrote:
"There IS a decoder that I would say is suitable for widespread
shipment. The mainline theora implementation has 3 major problems: 1)
it doesn't implement 100% of the spec. 2) It's not 100% robust against
invalid bitstreams; a maliciously crafted one could crash it. 3) it's
not very fast. derf's theora-exp implementation has an encoder that
doesn't work, but I assume that's irrelevent for a browser. The
_decoder_ fixes all three of those major issues. we haven't done a
formal release of that decoder, but it's in a releasable state - if
someone was seriously interested in giving it widespread use we'd do the
work to release it."
The real key is the format. not the decoder. If the format is
well-documented, open, non-proprietary, and high enough quality for the
web (for several definitions of "quality") then browser vendors can
write their own decoders, as open or closed source. It's nice if there's
an existing decoder they can just adopt, but it's by no means necessary,
especially for closed source vendors with the resources of Microsoft or
Apple.
The real question is whether such vendors would add support for such an
open format that competes directly with their own proprietary formats. I
suspect the answer is no, but I hope I'm wrong about that.
--
Elliotte Rusty Harold [EMAIL PROTECTED]
Java I/O 2nd Edition Just Published!
http://www.cafeaulait.org/books/javaio2/
http://www.amazon.com/exec/obidos/ISBN=0596527500/ref=nosim/cafeaulaitA/
