Indeed, various of our ML algorithms [4] and our matrix multiplication chain rewrite [8] are based on existing textbook algorithms. This means that we implemented these artifacts (loosely) based on the ideas or pseudo-code described in these references but never directly took over existing code. I'm no expert in licensing aspects, but intuitively, I'd say that the references are appropriate for both documentation and to give credit. Hence, I don't see a need to include them into LICENSE or NOTICE. @Luciano: Do you know of an ASF guideline on this matter?
Out of curiosity, I just double checked how Spark handles this since MLlib is in the same situation of relying on existing textbook algorithms. They similarly include references to papers at source code level and (as far I could see) only include originating source code / libraries into their LICENSE/NOTICE files. Regards, Matthias From: Deron Eriksson <deroneriks...@gmail.com> To: dev@systemml.incubator.apache.org Date: 05/12/2016 04:34 PM Subject: Citations Here are Justin Mcclean's finding from our last release (0.9.0): https://www.mail-archive.com/dev@systemml.incubator.apache.org/msg00482.html I think most issues raised by Justin have now been resolved. Could someone look at [4] and [8] to see if we need to cite anything in the LICENSE or NOTICE files or the Algorithms Reference bibliography? His questions #8 was WRT RewriteMatrixMultChainOptimization, which contains the following comment: /** * mmChainDP(): Core method to perform dynamic programming on a given array * of matrix dimensions. * * Thomas H. Cormen, Charles E. Leiserson, Ronald L. Rivest, Clifford Stein * Introduction to Algorithms, Third Edition, MIT Press, page 395. */ Thanks! Deron
