This Time Self-Hosted
dark mode light mode Search

FreeBSD, NIH and RPM

Okay, today I really feel pissed off about the continue rantings about Gentoo/FreeBSD and I want to state again what I think about this.

Gentoo/FreeBSD is a great opportunity for both FreeBSD developers and users. During all this time, ports are the only way to build software for FreeBSD, and this means that many people just assume that ports “will take care”, of patching, of fixing, of blacklisting. “If it’s not in ports, it won’t work.” It’s not true. Things can work if they are not in ports, just need to be fixed in case. Unieject is not in ports, but it works, for example.

Having to patch things “downstream” is a continue waste of time. Every time a new release of a patched software is done, the downstream maintainer has to check which patches applies, which are no more pertinent, and which needs to be redone by scratch. When patches are applied upstream, they can simply be dropped.

On Linux this practice is more followed because upstream releases for many different downstreams, and they all want to do the less work possible, and sometime they also like to share their fixes with everyone else who could need them. On FreeBSD, being ports the only way to install software (or this being what they want you to think), patches are not always sent upstream, making life difficult for users that does not want to use ports.

Gentoo/FreeBSD is the first try to get a different way to install software, not using ports, fixing things that needs to be fixed upstream. Is this reinventing the wheel? Yeah kind of. But improving it with more eyes. The more the merrier, the more the better. We can be of help as we can focus on things that ports does not focus on. We have no interest right now to support 4.x series or <5.4 versions, so we can focus on the part we take care of. When our base system is working fine and we’ll have enough people working on it, we could also focus on packages that ports ignores.

Unfortunately, while some devs and users seems to like the idea of a parallel work, also if they don’t want to switch and prefer their own system, some of them seems like suffering from NIH syndrome (Not Invented Here), trying to disturb us from our goals, and making our life harder. All I can say is a great “Thank You” to the devs that are willing to help and to accept our patches, The others.. we’ll see what will happen, and then talk about it, eh?

So what’s up with RPM then? Well I think that what make many people think that BSD is dead is because of the above mentioned problems. Having more than two teams working, focusing on different issues, is probably what made Linux so successful: there’s Mandriva for newbies, there’s Fedora for who wants, there’s Debian for who likes stability, there’s SuSE for enterprises, there’s Gentoo for who knows what he wants πŸ™‚ All of them contributed something to the general Linux landscape, all of them continue working on specific issues.
What I’d like to see now, is an RPM-based FreeBSD distribution. It can be more tricky than Gentoo/FreeBSD as at least ports and portage share the source thing, but I don’t think it would be too difficult to create a true FreeBSD (not a FreeBSD derived, of course) with RPM as software manager. If somebody is ever going to try to do this, I hereby promise that I’ll help with patches and with debugging of programs that need fixes.

Ok well ending here.. somebody did read my article?

Comments 3
  1. You wrote: What I’d like to see now, is an RPM-based FreeBSD distribution.It would help people to get accustomed to the idea that FreeBSD can be cool to install and use, but by itself FreeBSD does not need it, and RPM is an atrocious tool anyway.What instead I would like to see NOW is a major company backing an X11/Unix system with multimedia and gaming capabilities. No, MacOS-X is not Unix enough πŸ™‚

  2. Not like I love RPM, at the contrary πŸ˜› If I was not using gentoo I probably was on Debian, but RPM is a good way to give the idea I have in mind.RPM is simpler for many users, and it’s powerful enough to be used on FreeBSD.For what it concerns me, it can be DEB-based or YNI-based, it can use synaptic, yum or yast. I just would like to see a binary FreeBSD distribution, not using the tar.bz2, but more elaborate solutions and see what will come up from that.

  3. I would be an interesting thing to see for FreeBSD. RPM, might be a powerful base to start from… but my biggested concern (even as a Fedora user) is that the development of RPM mainly goes on in silent. If RPM changes, it would be something that would hit rather sudden, which would be another big problem if you ask me.Still, it would be a good way to just jump into the BSD’s.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.