[License-review] For Approval: OpenMDW License Agreement, versions 1.1 (OpenMDW-1.1)

[Luis Villa]

On Thu, Sep 10, 2026 at 6:57 PM Luis Villa <luis at lu.is> wrote:

> With the usual caveat that I’m just dipping my toe in this thread and
> reading every fifth post at best:
>
> On Thu, Sep 10, 2026 at 6:08 PM Pamela Chestek <pamela at chesteklegal.com>
> wrote:
>
>> What I find distinguishable about a termination of the copyright license
>> versus the patent license is that copyright infringement doesn't happen by
>> accident. There is no copyright infringement without deliberate, knowing
>> copying. Someone may have made the judgment call that the copying is
>> lawful, and therefore not an infringement, but it was nevertheless
>> intentional conduct.
>>
> This conflates two different things: copying is always intentional, but
> not all copying is copyright infringement. There are very clear good-faith
> arguments (under both the EU’s TDM exception and US fair use) that the
> sorts of copying we’re mostly talking about here are not copyright
> infringement.
>
> Now, maybe you don’t agree with those arguments (I move back and forth
> myself a lot of the time!) but intentional copying doesn’t mean intentional
> infringement.
>

I’d add that many of us here[1] strongly supported intentional copying that
was found to be fair use in Google-Oracle, even arguing in amici that
deliberate copying was a useful and important part of the software
development ecosystem.

(Less remembered detail of that case: Oracle’s original infringement claims
included an allegation that Google had copied code that was in fact PD, and
that both Sun and Apache/Android had copied from the same upstream
academic. “What if that 3d party PD code had this clause” is a thought
exeriment that probably cuts both ways, since (1) Sun almost certainly
wouldn’t have included the code if it had that clause; but (2) if they had
the dynamic of the litigation would have been very different and possibly
much shorter.)

I’m _not_ saying LLM training[2] is the same thing as an API
reimplementation; the nature of the copying and the industry power analyses
are very different. Just pointing out that copying sometimes has very
different valence in FOSS (which is why CC2+, GPL3, and MPL2 all explicitly
affirm fair use) and it’s worth remembering that before we make broad
statements about copying.

[1] Hi Jim, been too long ;)
[2] or RAG, which is also worth considering separately

>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opensource.org/pipermail/license-review_lists.opensource.org/attachments/20260910/db2b11db/attachment.htm>