When is a release not a release?

Off the Beat: Bruce Byfield's Blog

Sixteen months after its last release. OpenOffice.org has released version 3.4, its first as an Apache Incubator project. The release was covered matter of factly by The H (http://www.h-online.com/open/news/item/Apache-OpenOffice-3-4-0-debuts-1570353.html), and with a dash of skepticism by Brian Proffitt (http://www.h-online.com/open/news/item/Apache-OpenOffice-3-4-0-debuts-1570353.html). A week ago, it was even trash-talked by LibreOffice developer Michael Meeks (http://people.gnome.org/~michael/blog/2012-04-26-ooo-comparison.html), whose eagerness to discredit it was just a bit too obvious.

I can't help wondering, though, whether this latest version is a release in the ordinary sense of the word. That is, I suspect it isn't intended for end-users at all. Instead, it may be meant only for developers of related code bases.

Reading between the lines
I first heard broad hints of this possibility at the barcamp after ApacheCon in November 2011. I'm not sure if every developer is aware of it, but leaders of The Apache Foundation sounded aware of the possibility at least.

Consider: as dozens have pointed out today, OpenOffice.org is now far behind LibreOffice, not only in coding, but mind share. LibreOffice has a sixteen month head start, and the only major distribution still carrying OpenOffice.org is the notoriously outdated Debian stable.

Nor does OpenOffice.org have any real chance of becoming a serious competitor to LibreOffice. It appears to have far fewer developers, and, even more importantly, its code licensing allows LibreOffice to borrow from it, but doesn't allow it to borrow from LibreOffice.

All this is so obvious that you have to assume invincible ignorance on the part of The Apache Foundation in general and Apache OpenOffice.org in particular if you assume that the new release is actually meant to be competitive. Technically, that is possible, but, more probably, everyone is well aware that OpenOffice.org can't match LibreOffice.

If nothing else, that would explain the lackluster release notes (https://cwiki.apache.org/confluence/display/OOOUSERS/AOO+3.4+Release+Notes). Features like a revised color picker or a selection of line ends are welcome enough, but these are minor enhancements. Some new features, too, may be replacements for dubiously licensed pieces of code rather than outright improvement. By any standard, they are not features that would convince average users to switch back to OpenOffice.org.

This awareness would also explain the general sloppiness of the release. I mean, .rpms only at a time when Debian derivatives like Ubuntu and Linux Mint are the most popular releases? A failure to check package names to avoid conflicts with LibreOffice installations? Add inconsistent rebranding in readmes and release notes, and the appearance is of software either not ready or not intended for the desktop -- and, perhaps, of a shortage of workers.

By contrast, what's one of the first accomplishment mentioned in the release announcement? (http://www.openoffice.org/news/aoo34.html) The fact that, according to Jim Jagielski, ASF President and Apache OpenOffice.org has now been transformed "from a codebase of unknown Intellectual Property heritage to a vetted and Apache Licensed software suite."

By itself, that quote might be coincidence. Yet the same priority is given in the release notes. Under "New Features in Apache 3.4," the first point mentioned is that "The License has changed to our simpler, non-restrictive, Apache License 2. The Apache License will make adoption, distribution and modification of the software easier for all users and developers." For those who remember the KDE 4.0, citing developers as an audience can't help but seem a major indication of a release's direction.

True, the announcement and release notes also include suggestions that Apache OpenOffice.org is a project like any other one, and even the hope that it will soon become a top-level Apache project and move out of the incubation stage. But it sounds to me as though at least some people at Apache are of the opinion that OpenOffice.org's future lies in being a provider of code to other projects, rather than as an end-user application.

The future plans mentioned in the announcement are consistent with this goal. Most of the plans are vague: "support for additional native languages, enhanced accessibility, usability and performance improvements."

The closest thing to a concrete plan mentioned is "adopting new features and improvements from the upcoming IBM Lotus Symphony contribution" -- that is, presumably, doing a code audit on this fork and integrating it for use.

Maybe additional goals will emerge, to say nothing of bug fixes and minor enhancements. But, for now, the immediate plan appears to be to continue down the same path followed during the last sixteen months. That hardly seems the direction for a project with ambitions of becoming an alternative.

If the guess is right
If I am right, then the Apache Foundation is doing the free software community a genuine service. Somewhat more troubling is the possibility that some members of Apache OpenOffice.org could be laboring under the illusion that they are creating something much grander than project and Foundation leaders intend.

But all I can offer is supposition. The fact it seems more reasonable than assuming that Apache members are blind to what every keyboard pundit can see doesn't mean that I'm correct. Enough bad blood exists between certain OpenOffice.org and Libreoffice supporters that I can't entirely rule out spite as a motivation for doing something pointless.

So how will we know that my guess is right or wrong? One sign would be if Apache OpenOffice.org never leaves the incubator stage. However, that seems unlikely. I'm not familiar with the ins and outs of decision making at the Apache Foundation, but I imagine that denying full status to a project that has fulfilled all the requirements could be politically difficult.

The possibility also depends on how embarrassed Foundation leaders are about Oracle's unwanted gift of the OpenOffice.org code. They never say, but at times some of them do sound like they would prefer not to be involved.

A more likely sign will be what happens once code from various forks are audited and released. Will Apache OpenOffice.org continue beyond that point, tackling major improvements of the interface or adding major features or modules? Or will it simply shut shop and turn out the lights, having done what it was intended to do all along?

We'll have to wait and see.

Comments