[License-discuss] Open source software licenses and the OSD

Lawrence Rosen lrosen at rosenlaw.com
Thu Nov 8 04:19:53 UTC 2018

The initial debate between RMS and the OSI founders was some confusing nonsense over socialism vs. capitalism (in technical lingo as software license language rather than politics). I was initially confused. I found myself in the capitalist tribe at OSI, although my socialist parents would have been ashamed of me for validating developers' and companies' profit motives to support hackers. But I had a mortgage too, and I no longer worked at a university to support myself.


The FSF folks found more pleasure in the word "free," which scared many companies then. GPL was their main license model, and OSI had to accept it. Even if that meant squeezing and smushing the OSD language, GPL had to be "open source," and so it was.


Both OSI and FSF wanted to find other ways that could allow developers to make money. The LGPL was an FSF compromise. The Mozilla license and Apache license (first version), almost everyone except FSF could accept. Dual licensing and other novelty license models were created. We at OSI accepted them all (except for a few) and declared that all our licenses were for "open" software that companies could use for free. 


By that time, Bruce Perens was in a feud with other members of the OSI board and we couldn't rely on him to help us reword the OSD. We added OSD # 10 without his help.


The arguments today opining on various OSD provisions and the impacts of "copyleft" and "server-side" licenses are no longer interesting to me. I was pleased with that original OSI compromise about the essence of open source to use the software; copy it; create derivatives; distribute them; and read the source code. I even summarized those basic rules in the front cover art on my 2004 book (although RMS didn't like my Freedom # 5). By allowing all that shared software to be called "open source," and treating licenses as merely an agreement between licensor and licensee, then basic open source freedoms could be protected even while variety and experimentation were allowed in the software industry. 


GPL copyleft was then for me merely a license condition which any copyright owner could assert. Licensors also tried to assert side conditions about derivative works, corresponding source, private use, neon-sign-like attribution clauses, the obligations and responsibilities of licensors, etc., which were acceptable to OSI "as long as the OSD allowed it."


My own conclusion in the early 2000's was that the OSD was poorly drafted in some important respects. It was vague and uncertain, and the "basic open source freedoms" were being obscured. It appeared that the panoply of such licenses was confusing our audience. People were complaining about license incompatibility. Cacophony! License proliferation! Some licenses "went too far...."


Every one of these problems over the years has resulted in quibbling online about the obscure OSD language for those mostly obscure licenses. It tires me.


Instead, as long as the five basic freedoms on the cover of my book are protected, software will be open source enough for me. That is why I have proposed this common definition:


“Open source software” means software actually distributed to the public under software licenses that provide that every licensee is free to make copies of the software or derivative works thereof, to distribute them without payment of royalties or other consideration, and to access and use the complete source code of the software.




-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opensource.org/pipermail/license-discuss_lists.opensource.org/attachments/20181107/9171317b/attachment.html>

More information about the License-discuss mailing list