Talk:Technical Reference:State Definition

State Vector

The section on state vectors is totally wrong. After reverse-engineering some data files with the help of the BCI2000Viewer, it appears that the state vectors are actually encoded in reverse-big-endian. In other words, if smaller numbers represent lower order bits, then the state vector layout mentioned in this section would actually be encoded as follows:

State Vector Byte 1								State Vector Byte 2								State Vector Byte 3
0	1	2	3	4	5	6	7	0	1	2	3	4	5	6	7	0	1	2	3	4	5	6	7
Running	SourceTime																unused
0	0	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15

Which on the physical machine actually maps to:

State Vector Byte 1								State Vector Byte 2								State Vector Byte 3
7	6	5	4	3	2	1	0	7	6	5	4	3	2	1	0	7	6	5	4	3	2	1	0
SourceTime							Running	SourceTime								unused							SourceTime
6	5	4	3	2	1	0	0	14	13	12	11	10	9	8	7								15

Fixed section on state vector encoding

Thanks for pointing out the error. State vector encoding is actually little-endian (LSB first) rather than big-endian, and its layout on a little-endian machine is straightforward as a concatenation of little-endian encoded state values. I fixed the description accordingly.

Pseudocode reference

Do you think it would be helpful to readers to put (in pseudocode) a way to read the State Definition?

int ExtractStateValue(unsigned char[] vector, int byteoffset, int bitoffset, int numberofbits){

    int numberofbytes = numberofbits / 8 + 1;
    char bytes[numbytes] = vector[ byteoffset : byteoffset + numbytes ];

    int ret = bytes[0] >> bitoffset;
    for ( int i=1; i<numbytes; i++ ){
        ret += bytes[i] << ( 8 * i - bitoffset );
    }

    return ret % ( 1 << numbits );
}

This is not compilable C or C++ code, as far as I know. However, I think it might make things clearer.

Cstocks 21:15, 4 January 2010 (UTC)