For the texts, I would either use Huffman compression or LC. LC is a new compression algorithm being developed by Matti Jokela. It uses LZ77 compression and needs no 1K in memory, like Huffman. However, the compression ratio may be not as good as huffman. Depends on the usage. LC will be released soon as function file for Usgard 1.1 Andreas
application/ms-tnef