In his blog entry posted in October last year, OpenOffice.org contributor Michael Meeks included data and statistics. that he said underscored "a slow disengagement by Sun" and a "spectacular lack of growth" in the OpenOffice developer community.
Coining it a "dying horse" and "profoundly sick project", Meeks called for Sun to distant itself from OpenOffice and reduce its ownership of the codebase.
While contacted, Sun was unable to respond by press time.
In an e-mail interview with ZDNet Asia, Meeks acknowledged Sun as a significant contributor to OpenOffice, providing the largest number of developers to the software. However, he said there should be more large corporate contributors involved in the project.
"I believe the reason why other large companies find it hard to wholeheartedly commit to OpenOffice is precisely that Sun owns and controls the whole project," he noted. "This in turn makes it hard to attract a wide contributor base."
Corporations invest heavily in Linux, he said, which is far more successful at attracting developers than OpenOffice has. Meeks' day job sees him as lead of Novell's OpenOffice.org development team, but stressed that his views on OpenOffice do not reflect that of his employer's.
Meeks added: "There are lot of process problems [associated with OpenOffice], including over-burdensome requirements for even trivial code change inclusion, and a lack of external engagement. But, these are fairly fixable inside a broader community."
Code contribution not accurate indicator
David Mitchell, senior vice president of IT research at Ovum, disagreed that a lack of community contribution should be interpreted negatively, noting that the success of open source products depends on user adoption, not developer profile.
"Projects will have entirely different developer commit patterns, depending on their maturity. Mature projects are potentially always going to have less new code commits than fledgling ones," Mitchell explained in an e-mail interview. "A major restructuring of a faulty and poorly structured codebase would also produce lots of lines of new code, but would not necessarily reflect a solid and healthy product."
The OpenOffice codebase, he added, is already "pretty mature" and "quite stable".
Peter Cheng, founder and general manager of open source consulting company, TargetSource, agreed.
Disputing Meeks' observation that the software is a "dying horse", Cheng told ZDNet Asia that the codebase for OpenOffice is still growing, albeit at a slower rate.
Because the software has developed into a large and complex project, its codebase cannot be expected to expand at the same rate as when it was first launched, he said. Based in Beijing, Cheng is also ZDNet Asia's open source blogger.
He noted that OpenOffice is also supported by a community of 451 contributors.
According to Khairil Yusof, an open source software developer and consultant at Inigo Consulting, a shrinking pool of active contributors is unlikely to be a major concern for large business users.
Due to the nature of open source and open standards, Yusof said, companies can procure the services of independent developers or open source service providers, to work on OpenOffice and fix issues critical to their business. Based in Malaysia, Inigo provides consultancy services specializing in open source software.
"This can be done independently even to that of Sun or Novell, much like what Novell has done with its version of OpenOffice," he told ZDNet Asia in an e-mail.
Should Sun relinquish role?
According to Mitchell, whether Sun should relinquish its OpenOffice role is debatable because doing so may not necessarily change things for the community or the product.
"The strong development governance that comes from having a primary sponsor may well suffer if there is a more fragmented or dispersed community," the analyst said.
Meeks, however, believes the software would fare better without Sun in the reins.
He explained that the community council that oversees OpenOffice merely has an advisory role and owns no assets in the software.
"It all looks nice and open but it has no real authority. The current governance is fairly ossified, and it is clearly not developer-driven," he said. "Worse, under the hood, the only legal entity that owns all the interesting assets is Sun."
"[Because of that,] you can translate [questions like] 'why are people not investing in OpenOffice' into 'why are people not investing in Sun'--the answer to which is fairly obvious, I suggest, [because] it is a for-profit company," Meeks said.
"If we could recast OpenOffice as a foundation, with fair and representative governance by developers--then yes, I am certain the project would grow far more healthily without a Sun-veto at every step," he said.
Yusof also supported Meeks' call for Sun to give up ownership.
"As is often the case with other successful and active developer community-oriented projects, it is often best that management of the project be managed by an independent foundation," he explained.
The Apache Foundation, for example, is funded by market rivals such as IBM and Microsoft, he said. The Gnome Foundation is also co-operated by competitors Red Hat, Sun and Novell, all of which actively contribute toward the development of the software, Yusof added.
Meeks added that doing so would also provide a good opportunity for Sun because it would build more interest around OpenOffice, reduce the existing fragmentation, and build a larger, more successful OpenOffice. Sun would then be well-positioned to earn substantial support revenue from this potential development, he said.
Mitchell noted that OpenOffice 3.0 was a major release in 2008, and is likely to see an increase in adoption this year.
Vendors that have dedicated development work on OpenOffice will also help drive its adoption, Mitchell said, citing IBM's Lotus Symphony, which was developed on OpenOffice codebase, as an example. The analyst said 2009 will be "a good year" for OpenOffice in terms of growth in its user base.
"[However,] whether Sun can turn the increased adoption of the OpenOffice core into a strong revenue stream is another matter," he said. "Monetizing MySQL and OpenSolaris is likely to prove easier for Sun, as the demand for enterprise support agreements is likely to be higher."
Cheng said: "Sun is good at technology, but not at business."
He added that while the company has over the past years embraced its software business, specifically open source, as part of its strategy, Sun is inherently "still a hardware company". "All of the company's intentions are [focused] on the sale of its machines," he said.
"I think the most important thing for Sun, with regard to an OpenOffice strategy, is to determine how to balance the community and business," Cheng said, adding that the developer community should be given room to "handle all the things on its own".
"If you take a look at the Eclipse foundation, in the past five years, it has built a good community ecosystem around the platform. Everyone who participates in the community can benefit from it, not only individual, but also business unit.
"OpenOffice should rethink its position on the open source community, improve the governance structure of this community, get more contributors from enterprises as well as individuals to join the project," he said.
Keep developer community active
Maintaining a healthy ecosystem and an interactive community is critical in open source software development, Cheng noted. He added that greater innovation will be spurred when more people are involved in the development process.
On that point, Meeks was in agreement, noting that developers are the biggest challenge that the OpenOffice project needs to resolve.
"We need to up our game, take pride in our code, read more of it, fix, polish and improve. We need to be faster, smaller, cleaner almost everywhere in the codebase--there is a huge amount to do," he said. "Of course, from a feature perspective we need to keep moving to stay competitive, too."
"There is no lack of work to be done in OpenOffice. For example, the new 'Base' component needs a large amount of work to get to feature-parity with Access, and we need a Visio competitor, and so on," he said.
Meeks noted a contrast between community participation on the Linux kernel and OpenOffice. "Linus [Torvalds] has said many times that the Linux kernel is almost [complete and] uninteresting, and that people should focus higher up the stack--yet, he seems to attract far more contributors.
"OpenOffice clearly has lots of scope for improvement, and yet, attracts a fraction of the contributors so there is clearly a problem."
Licenses, however, could prove challenging. Cheng noted that OpenOffice is currently governed by too many different licenses, including General Public License 3, New BSD License and Mozilla Public License 1.0. "Anyone who wants to build an application based on OpenOffice, the legal issue will be too complex for them to handle," he said. "I think that's why the OpenOffice ecosystem is not running as well as that of Eclipse."
No comments:
Post a Comment