[ohf-licenses] ohf-licenses Digest, Vol 4, Issue 2

[Terry Hancock]

Okay, for starters, this IS a totally new concept, unrelated to the OHPL
draft I posted, right? (I'm assuming so).

Greg London wrote:
> I was wondering if the boundary layer for where
> the license kicks in could be defined for physical
> objects in terms of any contiguous connective
> volume, with no internal holes.

Um. Do you mean literally? Or is this in some kind of phase space?

> There must be a mathematical term that describes
> this sort of topology.

It's topologically a sphere, right? I don't remember any better term for
it. The term "convex hull" comes to mind, but I think that's from a
different branch of math. I am not a mathematician. (IANAM?)

> If Alice creates a mousetrap design and puts it under
> this license, then Bob could modify the existing parts
> and those would have to be covered by the license.
> If Bob adds a new part, it would not be covered
> by the license.

I'm not sure that's the intent here. I think we'd like that part to be
covered.

> For physical objects, I imagine each individual
> component is inside some sort of copyleft balloon.
> Any modification to those components is covered
> by the license. If anyone creates a new physical
> piece in addition to the originals, that is not
> covered by the copyleft aspect of the license.
> 
> Maybe this creates some loopholes, but I think
> it still stronger than using something like GPL
> for physical designs.

I don't see your idea yet.

> I'm still a little leary of these "atomic" terms
> creating some sort of unintentional loophole,
> or simply creating so much confusion on the part
> of users that they don't understand and don't
> use the license.

Hmm. It seemed very natural to me, because this is the way we normally
abstract designs anyway. Electronics engineers do not waste time on what
sort of packages the resistors they use are in when drawing a schematic
(technically they do when designing a circuitboard, but it's a fairly
trivial part of the design). You just go and look up standard parts in a
catalog.

IOW, discrete electronic components are to electronic designs what words
are to a writer -- they're beneath the creative threshold for the work.

OTOH, somebody could, in principle "create words". Then their "atoms"
are individual letters and not words.

A similar situation applies to a mechanical engineer working with
standard gears, screws, various engineering findings, and machine
linkages between them. Those elements aren't really part of the design,
they are simply supposed to do their job. So, e.g. it doesn't matter
whether the gear is "Delrin" or "Nylon", as long as it can take the
specified load.

And there are similar situations throughout engineering.

The only unintended consequence I can think of is when it's unclear
whether a design element is an integral part of the design or an atomic
element used in it. The atomic concept means we're "black-boxing" a part
of the design at some level, and it's not always going to be clear where
that level has to be.

This is more of an issue for large "sub-systems" (e.g. of a car or a
spacecraft), where the divisions may be very ad hoc, than for
well-defined domains like PCBs or RTL. Presumeably, though, if that's a
problem, the author can simply push the boundary out to the "whole
vehicle" and the atomic level down to the "circuit board" or "mechanical
component". This is why it's good that licensors can define their own
domains.

Likewise, we don't want to claim that every design that specs an open
hardware integrated circuit must be open hardware (most people would
probably regard that as "non-free" -- you ought to be able to use an
"open hardware chip" any way you want to).

In fact, I suspect that a court would impose limits like this even on
GPL-licensed works. The difference is that the OHPL is meant to
*specify* to the court exactly where these lines should be drawn,
instead of leaving it up to the court to figure it out.

One thing that bears remembering is that we don't actually include
"atomic" elements in our *Designs* -- they are included in the
*Products* as part of the manufacturing process. The distinction in the
*Design* is that atomic elements are merely *specified*, not *designed*
(we do not include the plans for a capacitor in a PCB design, we just
say "a 25pF capacitor goes here, and this is what it's connected to" --
we don't actually care what the capacitor is made of or what its
internal geometry is, as long is it is "25pF" (and also within certain
wattage and voltage ratings)).

Cheers,
Terry


-- 
Terry Hancock (hancock at AnansiSpaceworks.com)
Anansi Spaceworks http://www.AnansiSpaceworks.com