Re: New version of articulate available

[David Kastrup]

Peter Chubb <address@hidden> writes:

> Moving articualte functionality into a performer (which I think is the
> right approach long-term) isn't a copyright issue, because you can't
> just copy the code (scheme into C++?).

Scheme performers are desirable at one point of time, but their
structure would likely be different from yours.

> The *only* tricky bit, the only part of articulate that contains any
> substantial IP, is calculating trill timings. And it's a real hack
> full of mostly but not perfect heuristics, and should probably be
> redesigned anyway.

Heuristics are usually more or less mathematics and not subject to
copyright.  Copyright protects the expression of an idea, not the idea
itself.  So translating the program flow into another language is a
problem, writing down the heuristics and implementing it in the
appropriate manner for some language independently isn't (short of
patents).  Large companies needing to steer clear of problems let one
set of people read the problematic code and write out the principles and
specs, and another set write new code according to specs.  A so-called
"clean room implementation".  Not that we needed to get as absurd as
that.

-- 
David Kastrup