An explanation of the difficulty of solving license proliferation in one sentence

Joel West svosrp at
Thu Mar 10 05:28:54 UTC 2005

On 11:32 AM -0600 3/9/05, Fink, Martin R doth scribe:
>Next, if it's helpful
>I'd gladly spend some time going through a set of reasons why an
>infinite number of licenses will stifle open source utilization rather
>than foster it.

We have almost an infinite number of PC assembly companies, since the barriers to entry are trivial. Intel or AMD will sell their chips (or distribute through MiTAC/Synnex) to any whitebox vendor no matter how small. But (per IDC data) 39 million of the 58 million PCs sold in the US last year came from the top 5 companies.

Why do we want to knife baby licenses before they are born? Yes IBM would have been happier if Michael Dell had stayed in his dorm room, but what about consumers? Or Alan Greenspan?

HP and IBM can vote with their feet. They could post a list of 5 (or 10) licenses that they've seriously evaluated and agree with, and say that they will give preference to those licenses in any in-licensing or redistribution. Similarly, OSDL (or some other trade association) could commission attorneys to make a license-vs.-license compatibility matrix for 10 licenses that are legally distinct and well known (hire Larry, he's already started).

The act of posting more information about some licenses but not others (like Brian's "gold standard") would steer most use to these licenses unless there's an important reason to do something else. After all, the problem is not the number of licenses, but the number of licenses being used.

So a lot can be done without OSI changing its mission from promoting the OSD to regulating entry into the OSI-blessed list. Forrest is right -- if OSI no longer wants to take an inclusive view of the OSD, some other organization will.


Joel West, Research Director
Silicon Valley Open Source Research Project

More information about the License-discuss mailing list