[hpsdr] OHL v. NCL (was "WHY WHY")

[John Ackermann N8UR]

Hi Joseph --

I think it's important to understand one of the reasons why some of the 
HPSDR boards are NCL instead of OHL.  I'm not speaking for any 
developer, or for TAPR; but as the author of the OHL and NCL, I'm 
familiar with the issues involved.

The basic problem is that unlike software, each copy of hardware has a 
real cost -- I like to say, "electrons are free, but atoms are expensive."

Boards like those in the HPSDR stack typically require more than $1K of 
development cost for prototype boards and parts.  When it's time for 
production, quantity has a huge bearing on cost.  The price break at 
each quantity level is usually very significant, and this applies to the 
board, to the components, and to the assembly cost.

That means that someone who wants to manufacture these boards needs to 
(a) cover up-front engineering costs; (b) maximize volume to get the 
best quantity discounts; and (c) minimize the volume uncertainty, to 
reduce the risk of being left with unsold inventory.  Some sort of 
exclusivity during the initial production run is the usual way to 
accomplish these goals and control the financial risk.

For example, assume TAPR has ordered parts to make 500 units of a board 
based on a non-binding "expression of interest" (orders ahead of 
customer commitment are necessary because of long lead times on some 
components).  If someone comes along in the mean time and announces that 
they will offer the board as well, that could split the market.  Two bad 
things can happen as a result: (a) customers might pay more because 
neither vendor gets enough quantity for reasonable pricing -- or worse, 
neither vendor is able to deliver because the smaller volume makes the 
pricing impractical; and (b) TAPR could be left holding the bag for 
excess inventory.

Given the cost of some of the HPSDR projects, and ~500 unit quantities 
for most, there's a lot of money at stake.  As a result, it's only 
prudent to require some sort of protection against "cannibalizing" the 
market while that initial inventory is on the books.

The NCL is one way to accomplish this.  It allows publication of the 
design information to the community, while protecting the manufacturer 
against cannibalization.  (There are other methods, such as an OHL-based 
dual-licensing model, which could have the same effect; I personally 
don't think the model matters as much as achieving the goal.)

Nothing says that "NCL is forever."  My understanding -- and again, I am 
not speaking for any developer, or for TAPR -- is that at least some of 
the developers plan to convert their projects from NCL to OHL licensing 
once the initial TAPR production is sold.  (And, let me make it clear, 
developers are not *required* to bring HPSDR projects only to TAPR -- 
there's nothing saying TAPR is the sole HPSDR manufacturing partner.)

Of course, developers may choose to use the NCL for other reasons, 
perhaps simply to ensure that the project stays in the hands of the 
amateur community.  In the end, what the NCL really does is give the 
developer discretion in deciding to whom, if anyone, he wants to grant 
the rights for commercial production.  Given the financial 
considerations in open hardware production that just don't exist in the 
software world, that doesn't seem unreasonable to me.

I hope this provides at least some background in the issues around 
turning open hardware into real products that people can use.

73,
John
----
Joseph Teichman wrote:

> Also, on the hardware front, there really needs to be more emphasis on 
> the openness of the hardware. Not all of it is available under the OHL. 
> I would not really care to contribute to a project that is restricted 
> under the NCL. If I am going to contribute, I would want to only do it 
> to a project that is really open and can benefit everyone. Also, like I 
> mentioned in a previous post, if there were a full set of boards 
> available under the OHL, I think that there would be much 
> better availability, which would encourage and enable more people to 
> contribute and develop this project.


 1242134311.0