lzw.py

Chad Netzer cnetzer at mail.arc.nasa.gov
Thu Sep 30 20:29:16 EDT 1999


"Dinu C. Gherman" wrote:


Here is what the FSF says that Unisys is saying about their patent (ahem):

    No license or license fees are required for non-commercial, not-for-profit

    GIF-based applications or for non-commercial, not-for-profit GIF-freeware,

    so long as the LZW capability provided is only for GIF. However, a license
is
    required if freeware is incorporated into, or sold or distributed with a
commercial
    or for-profit product, introduced in 1995 [or later], or enhancements of
products
    that were introduced prior to 1995.

I got this at:

     http://www.gnu.org/philosophy/gif.html

There is more to it (IBM and Unisys both are involved, IBM licensing position
is
unclear).

The uncompressing part of the LZW algorithm is apparently NOT patented, BTW.

In any case, it is a bummer, but you can modify your algorithm and release it;
it
is mainly for learning anyway, right?  I wrote some nice coding algorithms in
school (arithmetic, huffman, lz77, each with various adaptability
enhancements)
and learned a lot.  The LZW algorithm is really no longer top-notch; computing

power is such that you can easily write algorithms that make assumptions of
bigger
resources that are probably not covered by the LZW patent (which dealt w/
tree-pruning and pointer rescaling, etc.)

I really should go back and play with these ideas w/ Python.  Dealing w/ C
made many
implementation details VERY cumbersome (especially given that I was exploring
adaptive models), but the Python abstractions may make things very elegant (if
a bit
slower :)

Chad Netzer
chad at vision.arc.nasa.gov






More information about the Python-list mailing list