receiving said unedited replica at a client associated with said contributing author;

forming a first revised replica by editing said unedited replica;

partitioning said unedited replica into a first plurality of segments, wherein a segment of said first plurality of segments is delimited by a delimiter;

sorting segments of said first plurality of segments into a binary tree based on a comparison operator;

partitioning said first revised replica into a second plurality of segments;

finding a best match for a segment of said second plurality of segments in said binary tree to form a block of at least one matching byte;

creating a difference file, said difference file comprising: a plurality of copy packets and a plurality of difference packets, wherein said copy packets indicating position and length of said matched block, and said difference packets consisting of stream of bytes flanked by said copy packets;

storing said difference file in memory at said client to be transmitted to the managing author;

transmitting said difference file to a server associated with said managing author;

reconstructing the first revised replica by finding in the original document said matched block using the memory position and length as indicated in the copy packets; and

forming a second revised replica identical to the first revised replica by copying said matched block into the second revised replica and inserting the difference packets into the second revised replica.