Yesterday's blog entry by Joe Brockmeier about why he found the comments reported to have been made by Sun officials to be petty politics has drawn a response from Sun's Simon Phipps as well as a clarification to a blog I wrote on the issue from Open Source Diva Danese Cooper (formerly of Sun, now with Intel). It has also flushed out some clarity over a connection that exists between IBM's Lotus Workplace and OpenOffice.org (OO.o).
Two comments, one by Sun's Phipps, the other by OO.o project leader Louis Suarez-Potts caught Brockmeier's eye. The Phipps comment that captured Brockmeier's attention came at the end of a negative report about OO.o by the Australian-based ComputerWorld Today's Rodney Gedda (see OpenOffice delayed thanks to lack of developers; Not all is rosy in the open-source garden). At the end of his report, in the context of Phipps' welcoming of contributions to the project from organizations and individuals, Gedda reported Phipps as saying "Ask IBM why it uses OpenOffice but doesn't contribute to it." The second comment -- made by Suarez-Potts -- can be found in a report by VNUNET.com's Tom Sanders. One need read no further than headline of that report -- OpenOffice team wants IBM contribution -- to get the gist.
Although Suarez-Potts isn't a Sun employee, Brockmeier connected the two dots, and characterized the connection as Sun-contrived "meme" that's "fishy" given the way Sun singled-out IBM and how Sun, because of the copyright assignment requirements connected with OO.o, could derive material benefit from such contribution (since when would IBM ever voluntarily do something that benefits Sun?). I glommed on to say how these "developments" also reveal the very nuanced complexities of open source licensing (including the lesser-understood benefits of copyright assignment) and Danese Cooper fixed my implication that those assigment requirements were a part of the licenses (the LGPL and SISSL) that go with OO.o. Mea Culpa. They aren't. According to Cooper, those requirements are a part of the "Joint Copyright Assignment, a separate document which must be signed and on file at Sun before a contribution can be accepted into the main OpenOffice.org source tree from outside of Sun." Document specifics aside, the copyright assignment requirement is factual and Cooper's explanation (beyond what I said) of why it's necessary just further highlights how complex open source licensing is.
But, in his blog, Phipps argues that, like in the children's game of "telephone," his words were misquoted, twisted, and taken out of their original context by the time Gedder and Brockmeier got through with them. Wrote Phipps:
"If Joe had bothered to ask, he would have discovered that my comment (which was actually a personal comment to Rodney Gedda over a beer and not intended for publication, but which I nonetheless stand behind once placed in context) did not relate to personal use of OpenOffice by IBM employees......the comment instead related to the common assumption in the OpenOffice.org community that IBM's Workplace Client product is based directly on source forked from OpenOffice.org about two years ago."
Phipps' response raised two obvious questions for me. First; Why, in common open source fashion, isn't IBM contributing code back to OO.o if it has a commercial derivative on the market? Second; WHAT!? Lotus Workplace is based in part on source code from OO.o?
Regarding the first question, the IBM spokesperson I spoke with -- Michael Shamrell -- is working on an official response. But based on what Phipps told me, IBM can elect to operate under either of the two licenses that go with OO.o and according to his understanding of the SISSL open source license, unlike some other open source licenses, IBM is under no obligation to contribute any modifications it makes to the OO.o source code -- even if it's for commercial distribution -- back to the project. So, Phipps' question was more philosophical than it was legal. Sort of like, if you're going to make money on this, then do the groovy thing by giving back to the project. Phipps also takes issue with Brockmeier's assertion that by being critical of IBM's actions, Sun is a pot calling the kettle black since Sun itself may have engaged in the same practice with projects like Samba. As he so often has done over the years, Phipps basically reasserts that Sun's record when it comes to contributing to various open source projects speaks for itself. My take is that this isn't about being critical of Sun's open source record as much as it is about how easy it is for anybody to "conveniently" pick projects in a way that can make any company look like a poor sport when it comes to open source. To be fair, both IBM and Sun have made significant contributions to many open source projects.
Then, there's the Lotus Workplace connection. As it turns out, Phipps is right and it's a fact that IBM has never intentionally hidden from the public. So, there's nothing nefarious afoot (nor was that ever implied). According to IBM's Shamrell, "Lotus Workplace's client tools -- the spreadsheet, word processor, and presentation software -- are based on OpenOffice and we disclosed that back when Workplace was announced in May 2004." While Shamrell works on finding out why IBM's code was never contributed back to OO.o, I think it's pretty obvious for the reasons already stated (Update: IBM issues official response). There's no obligation and just the same way Sun's application of the CDDL open source license to OpenSolaris prevents a competitor from profiting off a derivative of Sun's intellectual property, IBM can, should, and will exercise similar caution when it comes to how it works with open source code as well.
Tomorrow, I'll blog about what else I learned about Lotus Workplace in my conversations with Shamrell. It's pretty interesting stuff.