It’s been a little more than a week since Bryan announced Exherbo. The general interest was quite overwhelming and along with it came a lot of questions that had to be answered. I think most of it was covered quite decently but of course a certain amount of FUD was spread as well — both by people who should at least be somewhat informed and by people who didn’t quite get what we talked about. I’ll try to answer some of the common misunderstandings and hopefully help reduce the FUD a bit. I won’t delve into any technical aspects as Bryan, Ciaran and others are far more capable of explaining those.
If you want the short version of this, Mike Kelly already wrote it.
We don’t need yet another distribution
Quite the contrary. You might not need yet another distribution if you’ve already found one that suits you but those of us working on Exherbo definitely do. We want a distribution that fulfills the goals we stated on the website and to our knowledge, no other distribution does that. We’re not really interested in being bombarded with “Ah! But then please try distribution X – it will even cure cancer!”-messages at the moment, as we’re already past that stage. Even if no-one else finds our work useful it will still be useful to us and that’s what matters really. I hope that once we reach a somewhat stable stage, the result will be usable by others as well, but a closer study of our goals will show that “a billion users” and “a large thriving community” aren’t really something we aim towards.
Lastly, it’s also our hope that other distributions might benefit from the experiences we make. We are currently toying with a lot of different approaches to some common problems and hopefully some of our solutions will turn out to work better than what’s currently on the market. People are obviously welcome to look at what we’re doing and grab ideas from us (this being an Open Source world an’ all).
Oh, and a personal rant: Why is it the knee-jerk reaction to all new initiatives is “why do we need yet another”? There’s a reason we’re not still using Multics, there’s a reason we’re not still writing on stone tablets and that’s because someone was tired of carrying the stone tablets and thought he could do something smarter.
We don’t need yet another package manager/package format
This is mostly stated by people outside the gentoo community. Paludis has supported the gentoo ebuild format for a very long time and as such, Paludis has had lots of real world testing and is already technically superior to pretty much any other package mangler out there, at least in our opinion. Paludis is capable of supporting the rapid development process we’re currently in, where our exheres-format tends to change quite a bit as we introduce new features, deprecate others and modify the remaining lot just to see if we can do things even better than we currently are.
The exheres format isn’t really a package format in the way rpm, deb and the likes are. Exheres, like gentoo ebuilds, are advanced bash-scripts conforming to a certain EAPI that the package manager uses to download, configure, compile and install the packages. The exheres-0 EAPI is a fluent EAPI that will never be locked and can always change. But when at some point, we feel comfortable with what we got, we’ll create an exheres-1 EAPI that is locked and well-defined. This will allow other package managers to implement the EAPI if they want to and it will serve as a source of documentation for those writing exheres-1 exheres. This has another subtle but very nice side-effect: we can update our format without breaking existing implementation. A new exheres-2 EAPI could be made and rather than having old package managers choke on it and screw up systems, they can just bail out with a warning or error message. To gentoo people, this won’t sound too unfamiliar as the EAPI-concept is also somewhat working in ebuilds.
The ideas used the in the exheres-0 EAPI are based on some very smart peoples’ longtime experiences with gentoo and its shortcomings. So in conclusion, we’re not really trying to reinvent the wheel here, we’re trying to take a square wheel and turn it into an octogonal shape. It won’t run perfect right now, but at least it’s better than what’s currently out there. Remember that Exherbo is a source-based distribution so using debs, rpms and the like doesn’t really make sense.
We don’t need yet another init system
I’ll save the details and the fun for Bryan and his talk at FOSS Aalborg, but yes – we do need another init system. One that can accommodate modern computers and their needs. Check out Bryan’s blog after the 4th of June and judge his ideas there. Until then, kindly stfu.
Exherbo is all about pissing gentoo people off
Not at all! Pissing certain people off was a completely unintentional side effect.
In seriousness, we’re not really trying to make anyone mad, but if people take offense by the views we present then we can’t help it – that’s their problem and not ours.
Exherbo hates me! It says so on the website
No, what it says on the website is “It’s not that we hate you (unless we do)” which is quite the opposite. We don’t hate those who just want to try out Exherbo or those who are interested in the development of Exherbo. Nor do we hate those who doesn’t give a flying fuck about Exherbo. Those we do hate should be well aware of it and those who act like morons in our IRC-channels or on the mailing lists will swiftly receive some hate. But as the general rule of thumb you are not (explicitly) hated.
Exherbo isn’t a real distribution, you can’t install it yet
On the contrary! I’m writing this from Opera in Fluxbox on my Exherbo install. It’s true that we don’t provide installation mediums but those aren’t really required anyway. And Ciaran actually wrote a bit about installing Exherbo. But as we repeatedly state Exherbo isn’t fit for general use (yet) and we won’t support you. Questions on installing Exherbo will usually be ignored and in rare cases frowned upon or put up for public ridicule on our quotes page along with all the stupid things we’ve said :-).
That being said, we have a few useful packages in our repositories, but we’re still experimenting with more fundamental things at the moment, so it’s not a focus. Even once we reach a somewhat stable EAPI, it’s likely that we’ll hesitate to add a lot of generally useless things to our main repositories. We have no intention of growing an enormous developer base and adding tons of stuff only a few people will ever use. We have no intention of supporting 12000+ different packages but we want to make it easy to add new packages so that the general user and interested subgroups can maintain things they need themselves.
Exherbo isn’t fit for general use
Quite a few people have looked at our website and said this as if it was some huge revelation they reached after meticulously studying the content of the site. Allow me to quote: Exherbo is not, at the moment, a user-targeted distribution. There, it’s right there on the site, even on the front page. So how come some people see this as a well-hidden truth they’ve uncovered? It’s true — Exherbo isn’t fit for general use and that is why we want you to go and use ubuntu or gentoo or something else. It’s not because we hate you (unless we do) but we really have nothing to offer a general user right now and general users have nothing to offer us.