[geeks] Doorbells

[der Mouse]

> On the other hand, my understanding is that LZW is only an
> incremental development of Huffman coding,

If you think so, you need to read up on LZW and/or Huffman coding.
It's closer to an incremental development of Lempel-Ziv (which is the
LZ in LZW).

> Thought experiment:  How much could you NOT teach if you were not
> allowed to discuss or set assignments on anything covered by a
> patent?

You can discuss it all you like, in most cases.  What you are not
permitted to do is infringe it.  Discussing it normally does not do
this, though Bernstein mental process patents (LZW is one) are often
exceptions.

I do not believe it is possible to demonstrate LZW without infringing
its patent (well, except that (most?) LZW patents are now expired).  It
is probably possible to teach how to do it without infringing; reading
the patent, for example, will do for many people.  But building a
sample implementation *does* infringe.

This is one of the reasons I believe patent law is fundamentally
broken.  But it's still law.

/~\ The ASCII				der Mouse
\ / Ribbon Campaign
 X  Against HTML	       mouse at rodents.montreal.qc.ca
/ \ Email!	     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B