[M4IF Technotes] Reference software Version 2 cannot decode c omformance visual testing stream

[Rob Koenen]

Dan,
> Regarding the "rigorous interoperability testing in M4IF with 
> now over 30 participants in the 4th round, and cross-forum 
> testing with ISMA and others."
> mentioned by Rob Koenen in a near by post, will that result 
> in public code?

The main goal of the tests is getting interoperable *and* conformant
products (we are not satisfied with products that are all interoperable
but not conformant).
Interop testing does *not* test the reference source code, nor does it 
produce any source code. Interop testing also makes only limited use of the
conformance bitstreams, because the approach is to have n parties involved,
where approximately n parties encode content and approximately n parties 
decode that, so that we get O(n x (n-1)) results. (Not all parties do 
both encoding *and* decoding)
Whatever bugs are found in the standard are fed back into MPEG, and
this should in return result in improvements of the public code as well
as the bitstreams. Those bitstreams will be the basis for the certification
program that M4IF has currently under development.
Rob
ps: Forgive the plug - but those listeners who are interested in interop
    testing: please consider joining M4IF. We are a low-threshold forum, and
    joining is straightforward. Info is here: http://www.m4if.org/m4if/