"Copying"...what does that mean?

[Major Variola (ret)]

At 03:40 PM 11/21/05 -0500, Riad S. Wahby wrote:
>
>It's a violation of copyright to translate a book into a different
>language and sell it as your own, even if the two languages are
slightly
>at odds with regard to, e.g., their colloquialisms (i.e., the
>translation is "lossy" in some way).  It's about the chain of
>derivation, not the subjective experience.

No, its about the subjective (semantic) experience.  An MP3
is the "same", even if it is bitwise distinct.  Using a patented
algorithm is illegal (fnord), even if written in a distinct
language, for a different instruction set, and even if derived
independantly.  (Contrast with interfaces, which if implemented
in a clean-room situation, are not protectable.)  All this
under current US law, no endorsement implied.

Now making semantically distinct near-copies (aka, parodies)
are protected under the same rules, precisely because of the
semantic distinction.

Application to comp sci should be obvious.  An interesting
question is, say Bruce patented Blowfish, but you use e instead
of pi to seed the tables.    If he had patented the structure
and not the details, this would be enforceable; a question
remains as to how much you could vary the details of the
process and get away with it.

(Of course, a patented block cipher is about as useful as say
IDEA, heh..)