On Sun, 26 Sep 2010 17:00:23 -0700, Walter Bright wrote:
> The problem is, the BSD license *is* viral. If I look at BSD licensed > code, and someone accuses me of incorporating bits of it into Phobos, > then those bits must > be removed or Phobos becomes BSD licensed and so every user gets > infected with > it, too. > > If you say "that'll never happen", consider that twice that exact issue > has come up. > > Linking with a DLL is not viral, but statically linking a BSD licensed > library *is* viral. True. My "AFAIK" was wrong; I stand corrected: http://www.groklaw.net/article.php?story=20070114093427179 Summary: (a) the BSD appears to require that modifications be distributed only under the terms of the BSD, and that this requirement therefore cascades down to subsequent generations of code; (b) the license does not appear to permit the relicensing of BSD code under the terms of any other license, at least in so far as any restrictions in other licenses would seem not to be binding; (c) there may be some scope for arguing that the term “modification” to the code is restricted or limited in some fashion. However, as the license only permits redistribution of “modifications” the BSD does not permit the redistribution of any derivative work which is not a modification; (d) the BSD does not have a requirement for the distribution of source code. It is not clear whether this means there is a deficiency in the Open Source Definition.