27 Jul 2006

gentoo's public relations, community and innovation

Stuart's post about Gentoo's lack of organisation and ambition reads pretty much like the summary of the last 6 months of flamewars on gentoo-dev and gentoo-core mailing lists. I've read very little of any of the flamewars because I really just couldn't be bothered.

Stuart's observations are pretty much in line with what I see as well. There's been little innovation in Gentoo for the last two years, and it hasn't been able to pitch itself as a popular alternative Linux distribution that can compete with the binary ones around.

The PR Problem

Up until this month, the 12 months I've barely contributed anything to Gentoo. The main reason was the constant bickering (which has since subsided greatly) on the mailing lists. But apart from the flamewars on gentoo-dev, there was something else that was missing. I think Stuart is pretty close to the mark when he describes it as a PR problem, both internally and externally for the project.

The PR problem means innovative stuff in Gentoo does not get publicised internally, either because it gets bogged down by flamewars or people just don't know where to look. Developers do not know what is going on with other parts of the distribution. The solution here is not forcing each herd or team writing reports, but maybe a more encouraging attitude to blogging about new releases and cool stuff that you might be working on will be good. Things don't happen overnight, and I think initiatives like planet.gentoo.org is really helping communication both within Gentoo and outside of Gentoo. I personally learn more about the distribution from planet.gentoo.org than any other means.

One other thing that has annoyed me is the devmanual.gentoo.org (which I've only recently learned about), but really it needs a concerted push to get things up to date. I'm sure there are good graphic designers, web site designers in the community who wish to help, but don't know where to start.

Lack of innovation

Gentoo needs something more innovative, and something that engages people who want to contribute more. Instead of making the walls high to climb for people who want to make contributions, they should be lower and the development distributed.

That is why I reckon that the gradual acceptance of developer-orientated and user-contributed overlays will be the next "innovative" thing that Gentoo will see. I know, I know, all this overlay talk is just the same as apt repositories in Debian, but I think the official support for this model of development by Gentoo developers is a step in the right direction, bridging the gap between user submissions and hopefully be able to encourage people to share their contributions with other users more easily. Bugzilla really sucks for that.

Backing up a bit, I mentioned that this month I've started to contribute to Gentoo a bit more, because purely because I've been excited by a blog post (note, not the GWN) about overlays.gentoo.org. The basic idea is that developers and certain users can maintain an overlay which augments Gentoo's official portage tree. All this is managed by the nifty tool called app-portage/layman where you can select from a large number of overlays, for instance, gnome-cvs ebuilds, chinese support, etc.

For instance, my overlay right now consists of experiemental ebuilds for potentially portage breaking packages that need to be tested amongst developers, new software written by me in which I provide an overlay for interested people to subscribe to, and also user contributed ebuilds that I have poached from bugzilla so that I can test them and show the submitters that I am looking at their work and hopefully encouraging more interaction with them. Too long have good ebuilds languished in the Gentoo Bugzilla purely because no one has had the time/tools to test it.


The second thing that I have done is to start working on a set of tools to improve the development workflow for developers, and also users. I've written pybugz which is a Python command line interface to Gentoo Bugzilla (although I've also started using it on other bugzillas as well) to provide easy access for searching, downloading and closing bugs. This means a normal developer's job can now all be conducted from the command line. Using pybugz, I've probably increased my bug closing rate three-fold by not having to deal with Bugzilla's horrible web interface. This is available from my overlay, and will end up in portage as soon as I've deemed it stable.

Users can also benefit from this, if they are keen on reporting bugs, maybe they hit a emerge problem with a particular package, and instead of having to go to bugs.gentoo.org, they can quickly search with a single command "bugz search bluetooth" and find out all the bugs that may have been reported about bluetooth. And if they do not find such a bug, then "bugz post" will allow them to submit a bug report from the command line with out opening their browser, even submitting their "emerge --info" to apease certain devs!

"Turning the corner"

I ranted a couple of times privately to friends about wanting to quit Gentoo development, but I haven't because I still want to influence the distribution that I use on nearly any Linux box I touch. But in G. W. Bush's words, we might be "turning the corner."

Hopefully, we're at a start of a new phase in Gentoo development, and reconnect with the community. As to catching up to Debian, Fedora and Ubuntu, well, if we take care of the community and suck in more great people to contribute, then it will only be a matter of time.

You can reply to me about this on Twitter: