Subject: Re: compatibility and the OSD
From: Ernest Prabhakar <prabhaka@apple.com>
Date: Wed, 22 Sep 2004 14:09:45 -0700

 Wed, 22 Sep 2004 14:09:45 -0700
Hi Bob,

Bob wrote:
How is the Attribution Assurance License clause not a restriction?
It prohibits distribution of binaries that don't display an
attribution when *run*.


I don't remember the particular AAL discussion, so I can only share my 
experience as a programmer (caveat: I don't do programming for my 
employer, so this is a personal observation), elaborating on my prior 
post.

I am an Open Source programmer.I see some code which I want to use, 
which is available under an Open Source license.I download it, and make 
my modifications. Before redistributing, I go back to check the 
license. Oops, I have to make sure that I preserve some sort of 
attribution notice.I have to be a little creative, but I find some way 
to satisfy that requirement and still release my derivative work.I can 
slice, dice, extract, and reuse the code to the limits of my 
imagination, so at the end of the day I can deliver a product that does 
what *I* want --- as long as I pay appropriate homage.

Now I see your source code.I innocently download it and go through the 
same steps. But now, I discover that what I want to do is incompatible 
with passing your test suite.I want to do more, or less, or something 
quite different. I'm hosed.There's no way out.I can't deliver the 
product I want basedon your source code.No way, no how.

That violates my understanding of open source. 

Now, I appreciate your frustration; I personally consider the AAL a 
marginal license, in that it places what could be a burdensome 
restriction. But, at the end of the day the AAL allows me to create a 
derivative work which is redistributable under the terms of that same 
license. Preserving attribution can be an inconvenience, but its not a 
showstopper like conformance would be, which is probably why its not 
considered an OSD violation.

Ultimately I believe the OSD is about preserving a certain trust with 
the community. As a member of that community, I feel that branding a 
license like yours as Open Source would violate that trust.We can argue 
until we're blue in the face about what are legitimate interpretations 
of certain words, but ultimately the burden is on you to show that such 
a license complies with the spirit of the OSD -- most especially if the 
phrasing is ambiguous, as is arguably the case here.

Again, that doesn't necessarily represent official OSI policy, or the 
views of my employer, but that's how I see it.  Hope it helps.

-- Ernie P.
IANAL, TINLA, etc. etc.

On Sep 22, 2004, at 1:38 PM, Bob Scheifler wrote:

> You are not free to restrict what kinds of binaries can be distributed.
>
> How is the Attribution Assurance License clause not a restriction?
> It prohibits distribution of binaries that don't display an
> attribution when *run*.
>
>> You are not free to restrict what kinds of binaries can be 
>> distributed.
>
> How is the Attribution Assurance License clause not a restriction?
> It prohibits distribution of binaries that don't display an
> attribution when *run*.
>