 |
| Lawrence Joseph "Larry" Ellison (CEO OF ORACLE) |
Oracle announced a proposal this week to transfer the OpenOffice.org (OOo) project to the Apache Software Foundation (ASF). The move would put OOo under the umbrella of the Apache Incubator program and involve transitioning the project's source code to the permissive Apache License. The proposal is currently under review by the Apache Incubator Project Management Committee, which has not yet issued a decision.
The OOo project fell into Oracle's hands when the database giant acquired Sun. After taking control of the project, Oracle failed to address the long-standing grievances of OOo community members who were dissatisfied with Sun's mismanagement of the project. The problems—which related to governance issues, copyright assignment policies, and the difficulty of getting certain code contributions merged—festered under Oracle's brief stewardship of OOo.
The result was a major fork: prominent members of the OOo community established a vendor-neutral organization called The Document Foundation (TDF) to develop an OOo derivative called LibreOffice. TDF is broadly supported by the former OOo community and has the backing of many of the project's corporate stakeholders, including all of the major Linux distributors. Shortly after it was established, TDF reached out to Oracle and encouraged the company to participate in the effort and donate key intellectual property such as the OpenOffice.org name. Oracle responded with hostility and severed its ties with TDF supporters.
As independent OOo contributors flocked to the LibreOffice fork, Oracle was left without a robust community to help maintain OOo. Left with little choice, Oracle finally threw in the towel in April and announced that it would discontinue development of OOo and hand the project over to the community. LibreOffice supporters hoped that Oracle would take the opportunity to do the right thing and hand over the OOo project in entirety to TDF, but Oracle clearly has other ideas.
Dumping the largely abandoned husk of OOo into the Apache Incubator so that it can continue to be developed parallel to LibreOffice is not a particularly constructive maneuver. If Oracle had opted to take this route last year before its friction with the community necessitated the LibreOffice fork, it would likely have been welcomed by all parties. But handing the project to the ASF at this point, when a significant portion of the OOo community has already chosen to back TDF, is just petty and distasteful.
IBM, which remained on the sidelines when TDF's backers split from Oracle last year, has emerged as theonly independent proponent of Oracle's plan to give OOo to Apache. IBM says that it will assign its own developers to work on OOo through Apache and will commit resources to the effort. IBM's endorsement of the plan is disappointing, but it's not particularly surprising. IBM sells its own proprietary derivative of OOo and would obviously benefit if the project were to move forward under the more permissive licensing terms that ASF would adopt. The licensing change, however, could prove detrimental in the long term.
The Apache License is a permissive open source software license—meaning that it doesn't mandate the publication of source code changes. Permissive licenses are highly useful and are often a good fit for open source projects that are looking to enable broad commercial adoption and open the door for code to be purposed with the greatest possible flexibility.
Although I would normally applaud a company for adopting the Apache License and have used it myself on several occasions, the circumstances surrounding the Apache OOo proposal add a troubling dimension to the licensing change—especially if IBM ends up being the only major corporate contributor to Apache OOo.
OOo under the Apache License would give IBM the ability to decide on a case-by-case basis whether to open up and contribute new features or retain them as proprietary value-adds for its own commercial derivative. Over time, the open source OOo code base would become ghettoized and gradually fall further behind the proprietary versions.
OOo and LibreOffice are currently distributed under the GNU Lesser General Public License (LGPL), which was authored to support a "weak" copyleft model and differs from the GNU General Public License (GPL). The LGPL allows open source software components to be incorporated into proprietary derivatives, but requires modifications to the code itself to be published. It sets a good balance between permissive licenses and full copyleft licenses. This makes it a good fit for the open source office suite because it allows commercial derivatives while ensuring that major improvements to the core can still be taken upstream.
Brad Kuhn, the executive director of the Software Freedom Conservancy, offered some particularly cogent commentary about the licensing issue in his blog post about the Apache OOo proposal.
"It's my belief that this license change, if successful in its goals, may help foster a bit of a tragedy of the commons for the core codebase. The codebase is already well known for being somewhat unwieldy and time-consuming to learn. Those who take the time to learn it, but who aren't Free Software enthusiasts, may quickly decide that it's better for them to use that rare knowledge to proprietarize [sic] the codebase rather than contribute to the public Free Software versions," he wrote. "The LGPLv3 currently keeps such developers 'honest'; the Apache-License-2.0 will not."
TDF is also convinced that a copyleft license is the healthiest choice for the office suite, but is optimistic about the proposed licensing change because it will open previously closed StarOffice code and give TDF the flexibility to transition to a different weak copyleft license in the future.
"The Document Foundation believes that commercially-friendly, copy-left licensing provides the best path to constructive participation in, and growth of the project," wrote TDF's Italo Vignoli in a statement. "On the bright side, one benefit of [the proposal] is the potential for future-proof licensing. The Apache License is compatible with both the LGPLv3+ and MPL licenses, allowing TDF future flexibility to move the entire codebase, to MPLv2 or future LGPL license versions."
In the broader scheme of things, Vignoli says that the hand-off of OOo to the Apache Incubator would be "neutral" for TDF. The LibreOffice project would continue to move forward with its current governance structure and with the licensing terms of its own choice. The group is open to collaborating with the ASF to the extent that their interests are aligned.
If the proposal goes forward, I think it's likely that the Apache OOo project will have very little independent community engagement and will primarily serve as a venue for supporting collaboration between companies that are building commercial office software on top of OOo technology. LibreOffice, on the other hand, would retain the focus of the open source developer community and continue to move forward with delivering a downloadable user-centric office suite.
There is room for both, but licensing difference could prove problematic for cooperation between the two ecosystems. Improvements developed by the LibreOffice developers under the LGPL wouldn't be possible to merge into an Apache Licensed code base.
The ASF won't voice its position until the Apache Incubator Project Management Committee completes its review of the proposal. ASF president Jim Jagielski issued a positive (but noncommittal) statement on behalf of the organization.
"We welcome highly-focused, emerging projects from individual contributors, as well as those with robust developer communities, global user bases, and strong corporate backing," he wrote.
This parting shot from Oracle punctuates the company's legacy of bad stewardship and mishandling of OpenOffice.org. It's not clear yet whether the proposed Apache OOo will find its footing, but it seems likely that LibreOffice will continue to flourish as OOo's successor despite this move by Oracle and IBM to fragment the community.
Further reading
- Zdnet (zdnet.com)
- NetworkWorld (networkworld.com)
- Bob Sutor (sutor.com)
- Rob Weir (robweir.com)
0 comments:
Post a Comment