<div><div dir="auto"><span style="font-family:-apple-system,sans-serif">On Fri, Aug 28, 2026 at 3:10\u202fPM Richard Fontana via License-review <<a href="mailto:license-review@lists.opensource.org" target="_blank">license-review@lists.opensource.org</a>> wrote:</span><br></div></div><div><div dir="auto"><div class="gmail_quote"></div></div></div><div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex" dir="auto">On Fri, Aug 28, 2026 at 5:45\u202fPM McCoy Smith <mccoy@lexpan.law> wrote:<br>
><br>
> For AI licenses, OSI doesn't have looser or stricter standards; they are evaluated against both OSD <a href="https://opensource.org/osd" rel="noreferrer" target="_blank">https://opensource.org/osd</a> as with any license, and the OSAID <a href="https://opensource.org/ai/open-source-ai-definition" rel="noreferrer" target="_blank">https://opensource.org/ai/open-source-ai-definition</a> which is applicable to licenses designed to be used on AI. Beyond that, the evaluation is the same.<br>
><br>
> If, for whatever reason, people believe there should be other tests (like "advancing software freedom" or "Freedom Zero"), I'd be interested in hearing that (although several have already advanced those), but in order for the Board to understand what those tests are and how to apply them to licenses, I think we'd need to hear specificity on how to apply those tests in practice (both against the recently submitted AI licenses: ModelGo-Zero, ModelGo-Attribution, and OpenMDW-1.1, as well as against future licenses). I'd also like to understand if those tests are equally applicable to non-AI licenses, and if not, why not.<br>
<br>
Just to be clear, the OSI already is supposed to consider "software<br>
freedom", as I noted earlier, as stated at<br>
<a href="https://opensource.org/licenses/review-process" rel="noreferrer" target="_blank">https://opensource.org/licenses/review-process</a><br>
<br>
The purpose of OSI license review is, in part, to "Ensure approved<br>
licenses conform to the Open Source Definition *and* provide software<br>
freedom".<br>
(emphasis added - the important thing is that "conformance to the OSD"<br>
and "providing software freedom" are not equivalent)<br>
<br>
and<br>
<br>
"[The license-review participants'] consensus that a license does not<br>
ensure software freedom may, in some cases, be the justification for<br>
rejecting a license even where they cannot identify a specific aspect<br>
of the OSD or the approval guidelines below that is not met."<br>
<br>
That's not an articulated "test" but since adoption of that language<br>
in . . . 2019 or so? . . . the OSI has at least in some cases made<br>
reference to the concept of "software freedom" in explaining its<br>
dispositions of particular submitted licenses.</blockquote><div dir="auto"><br></div></div><div><div dir="auto">Yes, and as I\u2019ve pointed out repeatedly since 2019[1], the addition of this test with zero articulation of any definition of the term, or any tests, or anything other than \u201cwe know it when we see it\u201d invites exactly the sort of endless argument we see in this thread. </div><div dir="auto"><br></div><div dir="auto">Is CAL violating \u201csoftware freedom\u201d? no one can say, so we\u2019ll argue over it for a year. Is this license violating software freedom? no one knows what that means, so no one can say, so wellllp</div><div dir="auto"><br></div><div dir="auto">I actually do think values should form a part of the license-review process, and there\u2019s an interesting discussion to be had (maybe[2]) about some of what this license does. But \u201csoftware freedom\u201d isn\u2019t a value anyone at OSI can articulate (especially not \u201csoftware freedom, but different from what FSF means by it\u201d) so instead of discussing values we\u2019ll end up screaming about an ineffable and undefined term, hoping the current board agrees with us.</div><div dir="auto"><br></div><div dir="auto">[1] I see at least five posts on the topic in 2019 and 2020 before I gave up; those emails do not include archive URLs but I assume they\u2019re out there somewhere.</div><div dir="auto">[2] gotta say, seeing giant walls of text referring to \u201clebensraum\u201d with no intervention from the moderators is\u2026 not encouraging me to stay subscribed. </div>
</div>