Dr. Roy Schestowitz Latest posts | Real-time contact
Short bio: Software Engineer, interdisciplinary researcher, and an advocate of fair competition (read more)
Latest posts | Profile
Short bio: Computer Scientist, FOSS supporter (read more)
Tux Machines (TM)-specific
Submit search form
Many of you have dabbled in what's known as rolling release distros - PCLinuxOS, Arch... My suggestion is that Ubuntu adopt a hybrid release system. In Novemeber, Ubuntu could release a completely new release with full version numbering. Follow that up with the Spring release that's a rolling release of the November release, only completely updated with the newest Gnome, KDE, kernel, etc.. This way, the Spring release would be rock solid stable, yet have all the newest stuff, but the Novemeber release would allow a complete rebase and reassembly to incorporate new technologies such as Grub2, Plymouth, etc..
That is what they are doing already with the .04 release roughly corresponding in intent to the Mandriva spring release and the .10 release getting new features by the kilo.
But the real point is that it remains to see whether a problem actually exists. I mean Karmic is the fastest, most stable Linux OS I ever used since 2002 and the very first one with full hardware support for my laptop.
I don't mean to be pro-Ubuntu or something, I don't care, tomorrow I might switch to another distro, but my take on the big Karmic problem is that Ubuntu is getting more eyes upon than Linux ever did, many many more. So instead of 10 blogs about it you now get 1000, and you get everything, including the garbage, written by people clueless or with a hidden agenda.
I'm not saying let's reject criticism, but certainly pay attention to what we read, the bar speech style of many of those "reviewers" speaks for itself.
I understand your point, but let me make it a point to say, no, it's not what they are currently doing. Currently, both the .04 and .10 releases are built from the ground up - a total rebuild. With a rolling release, the second release would be binary compatible with the first because it would be nothing more than the first release, updated with a new kernel and software updates to that point in time. More or less, simply rebuilding the ISO with updates included. This would allow a new kernel for newer hardware support, as well as the ability for Ubuntu to keep hammering out the bugs and giving the users a full year support for a release. Those that have installed the first release could just keep up to date in the repositories and not have to install the new ISO.
I missed your point before. What you say is something they can maybe do after Lucid, when they will have pretty much defined their OS in full, then they can afford to sit back and cut their development speed by half, which is what your scheme inevitably involves.
More or less, yes. It also provides a longer term of support for the release, yet keeps the software up to date. In reality, it sucks to reinstall or cross your fingers and do a dist-upgrade every 6 months. If that could be put off to be once a year, but updating through the repository would give newer kernels, Gnome, KDE, OpenOffice and such. Also, the second ISO would allow new users and users with newer hardware to get the latest hardware support and have fewer updates to get after installation. IMO, this is a win-win situation: Less development costs for Ubuntu, more stable releases, longer support for releases, less frequent full upgrades, newest hardware support, newest software releases, etc... Yet, once a year, it gets completely rebuilt to incorporate new technologies such as Grub2 and Plymouth more efficiently.
Tux Machines is proud to be hosted by
Follow the site via RSS/Twitter.
Part of Bytes Media
Content (where original) is available under CC-BY-SA, copyrighted by original author/s.
Help Support TM
Wall of Appreciation
Powered by CentOS 6.5 (GNU/Linux), Varnish, and Drupal