🚀 go-pugleaf

In article <1991Sep11.183031.850@franklin.com> jesse@franklin.com (Jesse Perry) writes:
> What he actually argued was that it would be silly to patent
> an entire program, because such a patent would afford no
> protection to its inventor.  This is because people could
> change non-essential aspects of the program to create a new
> program with the same basic function, but without infringing
> the patent.

Sorry, but you're wrong. The LZW patent covers every aspect of, e.g.,
the ``compress'' program, including the output. It doesn't go into
detail on the methods of output, but there's no reason that it can't.
Claim 666: ``In the compression and decompression system of claim 1, any
method of providing the code signals using a standard input-output
library.''

Remember that patents often consist of tens or hundreds of claims.

> Therefore it would be useless to
> patent compress(1), because people could write equivalent
> LZW-based compression programs without stdio, which a patent
> on compress(1) would not cover.

Your error (well, really Zippy's error) is in assuming that
compress-without-stdio won't infringe on a patent just because the
patent covers compress-with-stdio. Any program which works like compress
will infringe claim 1, whether or not it infringes claim 666 (or any
actual claim in the patent).

> Dan> United States Patent 338029 covers not only software but
> Dan> an entire algorithm. It is clearly worded to cover that
> Dan> algorithm. It was granted in 1968.
> Is that number right?  Seems like maybe it needs another
> digit.

Whoops. That should be 3380029. Thanks.

---Dan