[Correcting] The Current State of Gentoo

My last blog post was the first one I’ve really just started writing on a feeling, and expressed it in full. Unfortunately, what came off in the post was all negative, without much hope for resolution, or even without anything positive about the projects mentioned.

So, here is a continuation of the last blog post I’ve written, in which I go over again what was bad, but also what was good, and how the bad can be corrected. Probably what the first post should’ve been.

The first thing I discussed was Gentoo Embedded, and its perceived “unfriendliness”. A big component of this feeling, I pointed out earlier, was a lack of documentation. This was too strong, and even wrong. The documentation does in fact cover most of the concepts around cross-compiling that’s important for getting started, and discusses how to build a cross-compiler from scratch. It fails to mention crossdev-wrappers, but it is not required for success. This is where the documentation ends, and leaves the reader to decide where to go from there. Most Gentoo users probably ought to know what to do to get the system they want, but it could be helpful to provide pointers from there. The documentation also doesn’t exactly flow from page to page, requiring a little bit of experimentation and extra thought.

Beyond documentation, though, there’s another problem that Gentoo Embedded deals with on a daily basis: cross-compiling, currently, is a huge topic in Linux all to itself that someone could make a career out of. There are so many ins and outs to choosing a cross-compiler, to picking USE flags, versions, and whatever else that it’s difficult to know it all. And knowledge is what is required to point someone in a particular direction without throwing them into a pit. (If I were to try and help someone, it’d probably end up being blind leading the blind, and all that). Only a few in #gentoo-embedded, not including me, have that kind of knowledge, and they aren’t always online or available to assist. (Maybe they’re all hiding in the mailing lists… I ought to sign up for those!) Those who are online are likely busy with other things, like work, and can’t get heavily involved or give heavily involved answers.

So what can be done to fix this? Well, it’s hard to correct the personnel issue immediately. In reality, it’s completely out of Gentoo Embedded’s control. However, the documentation can, and will, be corrected, if not by other volunteers who have already stepped forward, then by me (or, both). By idling on the channel, and paying attention to which questions go unanswered, or the general trend of questions, it should be possible to further customize the documentation in that direction, and further lighten the load on the channel.

To summarize, Gentoo Embedded covers a huge topic that is perhaps too much for the amount of knowledgeable people idling in the channel, but hopefully a change in documentation, which is coming, and a continual evolution to best fit the needs of Gentoo Embedded users, will solve all that. (This is where my discussion of how hard it is to change documentation comes in: compared to a wiki, the procedure required to get changes included is very difficult. That was the primary thrust of my previous discussion of Gentoo Embedded.)

The next part of my blog was about Gentoo Prefix. They discuss a much smaller topic, so their documentation was easier to write, and it was a lot easier to get into. However, it was when I started with Gentoo Prefix that I noticed a redundancy in Gentoo projects as a whole, and not just in Gentoo Embedded or Gentoo Prefix. This is where I dove into the flaws of the project system. Now, it’s good to note that I really like the way Gentoo sets up projects: if you have an idea for a big new feature or addition to Gentoo, you can make a project and develop it. You get your own CVS, web space, channel, mailing list, and so on, all within the Gentoo umbrella. This is beautiful.

So where does it go wrong? In a very subtle area: integration into official Gentoo. It seems (remember, I speak in opinions, not in facts) that many Gentoo projects are excluded from pushing their ideas back into main Gentoo for everyone who doesn’t already know about the project to enjoy. The benefits of integrating the mothership with its projects are: the Gentoo projects no longer has to reinvent the wheel, or constantly work around it; the Gentoo projects get recognition from the main Gentoo devs, who can then work with, rather than instead of, the projects.

How can we fix this? Get Gentoo developers to work with each-other more. Get people to meet and discuss how Gentoo can be better consolidated and made more streamlined. Really this is what I’m pushing for in a sentence: Streamline Gentoo.

This problem, perhaps, isn’t as widespread as I might think. Maybe it’s really just decisions high up for the betterment of the community. For example, they might’ve decided that, since 95% of the community doesn’t use it, we ought to leave it separate. The Gentoo documentation process (and thus the Wiki) could exist for the necessities of Q/A, issues that I’m not knowledgeable about.

I will continue to look around and see if I’m right or wrong, and what I can do to solve the issues I do find. I’ll keep writing.

Post a comment or leave a trackback: Trackback URL.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: