[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: My first message... more of a mad mans rant...

Steven Maddox (Cyorxamp) wrote:
Hi folks,

This is my first time on the Debian mailing list. I wanted to bring up a number of things I have been thinking about as of late. I had a long loooong chat in #debian-offtopic if some of you remember.

Abrotman - you're gonna love this :P

Apologies in advance for the anti-ubuntu-ness, it's just my opinion... please don't take it personally :D


First of all I want to bring up the idea of a FSF Debian GNU/Linux, everyone I have spoken to on IRC so far are in support of this idea so I thought I'd outline it here so that theres a record of it.

I'm tired of seeing projects like gNewSense and Gobuntu. Debian is miles more in tune with FSF expectations on 'the free operating system' idea than Ubuntu is, for example Firefox is a problem which Ubuntu don't mind but we do. It would take far less effort for us to create a FSF compliant operating system than it is out of a 'Ubuntu Base' <- (that makes my stomach churn). All we have to do is form an official sister project that makes a slightly tweaked version of stable (Etch) with a different Kernel that doesn't contain the binary firmware. Possibly if you wanted to get even more favour with the FSF you could swap IceWeasel for IceCat (since they are practically identical and serve the same purpose, but IceCat is GNU IceCat).

FSF endorsing an OS that has a 'Ubuntu Base' <- (again, *vomits*) only serves to give even more attention publicity to the Ubuntu community than to its daddy... Debian. Ubuntu being used as a 'base' for new operating systems (I can name many now, see wikipedia), or even as a 'Server OS' (are they Microsoft or what?) is a threat to what Debian is renounced for and could (one day) make us potentially defunct.

This sister project can act as a constant reminder we use binary firmware and keep us focussed on (one day, I admit its a long way off) moving towards that goal. And hopefully eventually (personal opinion here) GNU/Hurd... so it no longer matters anyway.

Debian has the upper hand in servers, stability, and multi architecture... it's time we show FSF why endorsing an OS that _only_ a x86 desktop-orientated OS is not enough.


NOTE: I only use Dreamlinux occasionally to see how its doing, I'm not a regular user and this isn't an advert - I'm using it as an example of a 'not so bad fork'.

Second! I spent a long time in #debian-offtopic (points at abrotman) arguing that Dreamlinux is a worthy Operating System. And the reason why I say it is 'worthy' is they (for the most part) actually use our repositories. If we were to declare lenny as stable tomorrow, all of their users would immediately start trying to upgrade to Sid (which is a little silly in my opinion, but hey).

I was trying to say that Ubuntu is a fork in that they have wholly taken a separate copy of Debian and re-branded it their own. The Ubuntu front-page doesn't even mention it is based on Debian any more now, although the server page does say... "The Ubuntu Server Edition - built on the solid foundation of Debian... blah blah"... but that does make me scream... WELL USE DEBIAN THEN. But that's a little off topic :P.

So Ubuntu is a definite fork... but I was trying to say Dreamlinux is at least _less_ of a fork in that it is very very much reliant on us. If you want to fix something in Dreamlinux there's a 99% chance that 'the something' is actually a Debian component... and to fix it you would be fixing it for Debian too as we host it.

I will admit... Dreamlinux IS A FORK... and it is not Debian (only based on) -- so you can be happy abrotman :D rest well...

I would like it to be mentioned on www.debian.org/misc/children-distros as I think it is an admirable project and is now in the top 10 of distrowatch. Also I think Sidux works in a similar way by using our real repositories (I like calling these Sporks not Forks :P but I'm just odd) and that isn't on the list either. Dreamlinux devs and admins make a repeated habit of saying they are Debian based, and they do not ever intend to go with their own repository - the developers are proud Debian users looking to make something preconfigured for desktop use out of the bo...ummmm .ISO :D

To some degree these Sporks (less than forks) make me think a win for them, is at the very least a small win for us... if we can have official KDE leaning and XFCE leaning editions of Debian (alternate CD1's) then these are like unofficial editions of debian with their own cause... but not straying totally on two feet from our project like Ubuntu.


Now for my final rant...
I've been using testing (Lenny) at work for a number of months now, I know its not meant for production environments but I get by. There's a few packages (e.g. grsync) that I like to use and I need their latest versions to get the features/fixes I need. Backport doesn't really have much in it... I got Pidgin out of it one day, then wondered what else was in there and found out not much was. Before I continue I would _really_ like to point out (because I'm new you may not know) I have to up more respect for this project and the length of time needed to create a truly stable operating system.

Every day there's a good twenty odd updates which makes sense since it is -testing-. A few months back I couldn't even boot up because lilo had updated and screwed the boot up process (don't say use grub, there's a good reason why I use lilo) but I fixed it, woo! I don't really mind the fixing of stuff and that was barely a problem (only a few hours :S). At the moment I can't use my number pad keys and every now and then focus gets stuck on an individual object (like a text box) in a program and you can't click or do anything else _on_ anything else. This occurred a could of week back... _after_ updates ;P DUN DUN DUUUNNNNNN.

I don't know why it happens and I'm not asking for ideas on fixing them... seriously don't try as I don't care. My suggestion however is this... In the process of lenny being created it can't possibly go from 0% done straight to 100% done. If we had Alpha releases, say 'lenny-alpha1' release at a point where there's no major block/crash-like problems being caused then people could download that milestone release. 'lenny-alpha1' could then not update until 'lenny-alpha2' is released... I'm not saying change how the testing development process already works, just have these additional milestones for people who want to try it in a production environment as a stop gap until lenny is ready. Essentially it would be something 'almost-stable'... 'testing' can go on adding more unstable packages after an Alpha release without affecting those users who liked being on 'almost-stable'.

Without this 'stop gap' I feel a lot of people probably just get frustrated and hop to Ubuntu with its 6 monthly release cycle. Many many other projects have alpha releases, I think this could work for Debian too and would keep a lot of people happy while we wait for the next stable release... which should _not_ be speeded up under any pressure - it takes... as long as it takes. These milestones don't have to affect the way it is developed, just someone keeping an eagle eye out for a weekly generated ISO that -kinda works OK- that can be relabelled Alpha 1, 2, 3 etc... and with a tweaked apt so it doesn't update until the next Alpha.


Please let me know what you think about all this stuff :D

I'm hoping your all friendly and don't bite.


Welcome to the club. I used to feel the same way when I first found debian (ubuntu showed it to me). I don't worry about it too much any more. I like how it will keep all the newbies away from us so I don't have to spend all my time running from several questions of "how do I navigate the cli" type stuff. I like testing. I have not had a problem with it. I don't think you could really do alpha style releases with it though. Testing is nothing more than packages that are 10 days old or so from sid that flow into it. Then one day they do a freeze and stop the flow of packages. Fix the bugs in the current packages and release the product. That is it. I am never going to change my source.lst from testing. It just always updates and stay with the times. No different stages are needed for me ;)


Reply to: