Three Forks in the Road for FOSS

September was an unusually exciting month here in the Linux blogosphere, not least because we saw the launch of three — count ’em, three! — forks, meaning that we now have three new software projects to discuss and watch.

First, of course, it was the creation of OpenIndiana to replace OpenSolaris. Next came Mageia, an alternative to Mandriva Linux. Then, just last week, it was the OpenOffice.org community’s LibreOffice announcement.

Special thanks to Oracle, in particular!

It was the Java-brandishing one, of course, that ultimately led to the creation of OpenIndiana and LibreOffice, while the mess at Mandriva gave rise to Mageia.

‘Goodbye Oracle, Hello LibreOffice’

It’s been a lot for Linux bloggers to keep track of, it must be said; said bloggers, meanwhile, had plenty of other ideas as to other things that must be said as well.

“Ever since I saw Oracle’s name written across OpenOffice, I have winced and felt very uneasy about exactly what Oracle’s version of OpenOffice was doing on my computer,” Ridcully wrote on LXer, for example. “Yesterday, OpenOffice 3.2.1 (Oracle) was removed from my main work computer and OpenOffice 3.2.0 (Sun) replaced it.

“That is now where I shall stay until LibreOffice releases its next full upgrade, and then I shall be staying with whatever the Document Foundation provides,” Ridcully added. “Like more and more people are saying: ‘Goodbye Oracle, hello LibreOffice.'”

‘This Is Great’

Open source forks tend to be a good thing, Slashdot blogger rtfa-troll pointed out.

“From the user point of view this is great,” rtfa-troll explained. “You don’t get data lock-in because the source code always lets you see how the formats work; you do get much faster advancing software and it doesn’t even really matter which fork you pick (though going with the community rather than the company has always been a good pick; just beware that often the community is with the company).

“Just because forks are bad in proprietary software doesn’t mean the same here,” rtfa-troll concluded.

‘Like Naming Your Dog BarkingMammal’

On a more superficial level, “Why do OS developers and other free software creators always pick user-unfriendly names?” Slashdot blogger future assassin asked, referring to Mageia. “Whenever someone who knows nothing about free software/linux asks me what free alternatives they could use, I get a weird look from them when I tell them about Thunderbird, Firefox, Ubuntu, Amarok, Gimp and etc…”

Blogger MBGMorden had an answer to that: “Yes because Excel, Powerpoint, Quicken, Maya, and Twitter are so much better . . .”

Then again, “that other Mandrake fork, PCLinuxOS, is just as bad a name,” haruchai pointed out. “Why didn’t Bill Reynolds call it Texstar Linux? Calling a distro PCLinuxOS is like naming your dog BarkingMammal.”

The analyses and one-liners have been flowing fast and furious for weeks, so Linux Girl knew it was time to dig deeper.

‘A Huge Amount of Work’

“Forks happen for a variety of reasons,” Chris Travers, a Slashdot blogger who works on the LedgerSMB project, told Linux Girl. “One of the major reasons — and the reason LedgerSMB broke from SQL-Ledger — is that the forking committee may simply lack confidence in the maintainers of the software.”

Forks are “a huge amount of work,” Travers explained. “They divide resources, and usually the forking project starts of at a disadvantage because most people do not want to quickly adopt the new fork. Forks can only earn success through a great deal of long-term effort, community building and so forth.”

It’s too soon to tell if the recent round of forks will go anywhere, “but we shouldn’t rush to judgment, either,” he concluded.

‘Stick a Fork in It, It’s Done’

LibreOffice and Mageia “are examples of two extremes,” said Barbara Hudson, a blogger on Slashdot who goes by “Tom” on the site.

“Mageia, the linux distro formerly known as Mandriva, the linux distro formerly known as Mandrake, is as dead as Mandriva,” Hudson opined. “With so many other healthy distros around, this is more a case of ‘stick a fork in it, it’s done.’ The ‘Mandrake Magic’ has been gone for quite some time.”

At the other extreme, however, sits LibreOffice — “a fork of a project that has been increasing its market share, and the No. 1 free office suite worldwide,” she pointed out. “This fork has broad support inside the community, including all the largest distros.

“Redhat, OpenSuse/Novell, Ubuntu, Google, FSF, OSI… it’s a safe bet that LibreOffice will be shipping in most distros,” Hudson predicted.

‘Sun Did a Poor Job’

“I suspect it’s far too late to make Mandriva anywhere near relevant,” agreed Montreal consultant and Slashdot blogger Gerhard Mack.

Nevertheless, “the OpenOffice fork will be a good thing,” Mack asserted. “Sun did a poor job of managing the project by refusing too many patches from external sources.”

Similarly, “I’m really excited to see LibreOffice spin up and announce that it’s taking on the features of the Go-oo fork,” Hyperlogos blogger Martin Espinoza agreed. “Sun kept OpenOffice development down for too long.

“Indeed, I long since stopped seeing the point in having it installed, and switched to AbiWord and Gnumeric,” Espinoza added. “If I’m going to have a dearth of features, I might as well forgo the bloat, too. Oracle was guaranteed to treat it even worse than Sun.”

‘Plans for a Full Stack Approach’

Oracle “simply doesn’t seem all that interested in playing the FOSS game, and I think I know why,” Slashdot blogger hairyfeet said. “I believe ultimately Ellison has plans for a ‘full stack’ approach, and Linux will NOT play a part of this.”

Rather, “it will be Sun hardware with a custom Solaris OS designed for said hardware and more importantly designed to give maximum throughput to the Oracle database running on top,” hairyfeet explained. “This is smart, as it gives ONE vendor for corps to deal with, and — unlike Linux — Ellison owns the stacks and thus can direct its R&D.”

The world “revolves around money, whether FOSS advocates like it or not,” hairyfeet added. “It costs money to keep up with MSFT and Apple when it comes to R&D. It costs money to pay for developers and get them to focus on the dirty jobs like bug fixing. This is why all volunteer developers simply won’t work in the end.

“How many bugs are currently open at Ubuntu, even with Shuttleworth’s money?” he pointed out. “Bug fixing is a lousy, hard and thankless job that, frankly, without being paid for, simply won’t get done. As we saw with Mandriva, these companies slowly sink farther and farther until they die or, like Sun, get bought for the IP.”

‘Too Occupied Suing Google’

Blogger Robert Pogson saw it differently.

“LibreOffice is the result of Oracle’s (or Larry’s) failure to be open about FLOSS strategy,” Pogson suggested. “FLOSS being distributed is about sharing.”

The OpenOffice.org community “saw how OpenSolaris was left to wither and took pre-emptive measures,” he noted. “The steps needed to be taken anyway to broaden participation in the project. There was not a large enough group of developers contributing to be healthy.”

Separating the code from Oracle “was the best way to open it up to participation by other large businesses who could use OpenOffice/LibreOffice as a tool or a product to distribute and could contribute developers and funds rather than paying licensing fees,” Pogson added. “That’s how Sun got involved years ago, but Oracle seemed not to be engaged.

“Perhaps they were too occupied suing Google to pay attention to one of the killer apps of GNU/Linux,” he suggested. “A bird in the hand has to be set free to make the world a better place. It is unnatural to keep it too close.”