Should MEPIS change to the Ubuntu kernel?

"If somebody can just clarify the issue of it still running ANY Debian repo package, I think the huge majority of the MEPIS user community will know how to vote."

That's exactly what I want to know.

--
Post questions on www.mepislovers.org too.
Check out our wiki: www.mepislovers-wiki.org

My sentiments too

What makes a kernel a "debian" kernel? The DCC kernel is modified, but still runs debian packages, where as the Ubuntu kernel does not. Is that what makes one kernel a Debian kernel and one not? Will using the Ubuntu kernel source make it not compatible with Debian packges? It seems that the question is to use the source, not the actual Ubuntu kernel as compiled by the Ubuntu people. If the kernel can be tweaked to still be compatible to the Debian packges, then go for it. My Ubuntu experience is that their universe packges are buggy, so using the main Debian repos is important to me.

As far as I know (I'm a newbie in this) the only packages that depend on the kernels are the drivers, I compiled vanila Linux kernels and used kernels from other distros without having to change any packages.

Then why would Mepis have any problems using Debian packages? I don't see the logic.

--
Post questions on www.mepislovers.org too.
Check out our wiki: www.mepislovers-wiki.org

Try using a kanotix kernel (debian sources I think) with Mepis?
Doesn't work.
Why?
A vanilla kernel will work with anything.
Why?
Will a present state Ubuntu kernel work with Mepis?
Why (not)?
I'm really trying to understand this.
anticapitalista

Philosophers have interpreted the world in many ways; the point is to change it.

"Try using a kanotix kernel (debian sources I think) with Mepis?
Doesn't work.
Why?"

Why do you know it won't? Because you tried it and didn't work, maybe you just didn't load the right drivers or didn't configure it right.

How do you think you are able to use RPM packages? Those were not compiled for Debian kernels.

--
Post questions on www.mepislovers.org too.
Check out our wiki: www.mepislovers-wiki.org

I read several articles on the internet that suggest that since early january, DCC, Debian and Ubuntu are in fact co-operating together to build a better linux. This sounds much different than DCC simply using Ubuntu's kernel.
This particular post comes from Xandros (DCC member) but it would certainly allay fears of Mepis users if Warren would elaborate further than the surprise announcement that started this all.
If, in fact, there is co-operation then we could expect that this would provide far more resources to keeping Debian on a scheduled release cycle, and all Debian based linux would benefit.
Google DCC ubuntu.
One post:
http://lists.dccalliance.org/pipermail/dcc-devel/2006-January/000487.html

Since this is dated January 5, 2006, it appears that it is already a done deal. Why the poll?

It seems to me, if Mepis is committed to the DCC alliance, and the DCC has decided to base the next version on Ubuntu sources (not a straight Ubuntu clone that requires Ubuntu repositories, mind you), then the decision seems straight forward; If Mepis is based on the DCC kernel when it comes out, it will get this anyway. It's inevitable and presumably for the better. So if Mepis preemptsthis, fine by me. It doesn't necessarily mean abandoning Debian compatibility.
A DCC member could, like Xandros, remain largely incompatible for their own commercial reasons, but for a small free distribution like Mepis, it makes no sense to break compatibility.
A major change, like the one with udev, usually has issues that need ironing out. So this change may have even bigger issues in the short term. But these issues, like going to X.org from Xfree86.org, or moving to GCC4, have to be dealt with at some point anyway.
So what's the problem?

EDIT: Having read the post above on the DCC mailing list, it appears they are working on a joint Ubuntu-DCC-Debian kernel to satisfy all concerned. They're working with Ubuntu especially with the kernel, Xorg, and laptops. Warren is apparently their Kernel compiler, so he's best placed to know the current status. It seems more a political issue, whether using Ubuntu sources is "bad" in some way as opposed to pure Debian.
There is talk of making DCC a Debian sub-project, and as of Debian Etch, it should be LSB compliant. Debian is a conservative slow-moving distribution, but by Etch, the differences between the Debian Stable and DCC kernel is said to be small. Whereas Ubuntu is not a full DCC member, they are happy to collaborate on specifics like the above, in particular the kernel. So I still see no problem. According to the mailing list, the DCC 3.1 spec is due mid-February, so there need be no rush to preempt it, even if it meant delaying the next Mepis release until the exact nature of the kernel is known (such as whether Ubuntu sources are used), which Warren is in a good position to know beforehand. I for one have no issue with Ubuntu, they've provided a lot of needed resources to Debian and contribute back to it.

Odds are you folks are already are on top of this, but when Warren upgrades the kernel, it needs dazuko compiled in it so that KlamAV can auto-scan.

Loye Young
CMCMaintenance.com

Doesn't politics get everywhere?? Well if the ubuntu kernel is better than the standard Debian one and it encourages Ubuntu to join the DCC then why not, Warren can always revert back on the next version, but if Mr Shuttleworth is going to continue on a slightly different track and is not able to see the wider picture then no.
James B

I believe this quote from a previously posted link really explains the politics and debat. Shuttleworth( Ubuntu ):

Shuttleworth said, "The DCC distro doesn't use the Debian kernel, and it modifies key pieces of the infrastructure like the linking system and core system libraries. So it's not really Debian at heart."

Ian Murdock, the founder of Debian and now a leader of the DCC Alliance disagrees with this assessment of the DCC common core.

"We went to great lengths to not modify the linking system and core system libraries, and what we came up with is an add-on, not a replacement, for what ships with Debian proper," said Murdock in an email interview.

MEPIS has already released a DCC Alliance core/Debian Sarge-based distribution, MEPISLite.

Shuttleworth also believes that the DCC distro is really aimed at shoehorning an LSB-compatible environment on top of Sarge. I think if LSB was a goal for Sarge then that could have been achieved directly in Sarge, not in this hybrid fashion. If we were to do LSB for Ubuntu, it would be done directly rather than as a compatibility layer."

That said, Shuttleworth finds that the "vision behind DCC, which is indeed compelling, is that it would provide a common platform for certification, and that the distros that make up the DCC would all ship exactly that same core."

"But it strikes me that this approach has never worked in the past." UnitedLinux, which briefly allied The SCO Group Inc., SUSE, Conectiva S.A (now part of Mandriva), and Turbolinux.

Beats me. Warren's child. Warren's decision.

I also found this at Xandros:

Hi Everyone,
A while back I proposed the idea of collaborating with Ubuntu to jointly
(Ubuntu and DCCA) maintain and work from the same kernel source tree.
This was well on this list received so we moved forward with the idea
and approached Ubuntu. They have agreed that this would be a good thing
to do and have begun setting up a process to work together. As for
laptops we're also looking at how to collaborate and jointly help
improve support for everything from keypad buttons to power management
(ACPI).

Recently we at Xandros have been in the process of putting together a
2.6.15 kernel tree for all the DCCA to use, this was at a very early
stage. We will now instead work from the Ubuntu tree and submit patches
to them for inclusion. From that tree the DCCA will grab source and
build it's own kernel binaries and drivers. The laptop project
development process is being worked out, and I'll let everyone know as
soon as that's setup.

Again let me reiterate, this is about what's best for the end user and
what the Linux community as a whole needs to do to get this done.
This is not about Ubuntu being a part of the DCCA, nor is it about the
DCCA being absorbed into Ubuntu, this is about two different communities
working together to help our end users and customers.

Thanks
Dave

Hi;
I've used 5.04 and 5.10 of ubuntu an they were ok.
But Mepis worked better to me.
But if this helps us get full 64 bit support and better hardware support as well as gets us more hardware to use /support it's a bonus to us and linux as a whole.
I myself would like to see a generic kernel used while the system builds a amd or intel 64,32,etc kernel in it'time of no use.
Will it happen um not as yet has it done so.

But thenm again I'd like linux to be widly supported too and get all the games and hardware I'd like to use too.

I've tried to load ubuntu 64bit 5.04 on a 64 bit laptop and had it lock up so 64 but needs help as well.
I'd like speed ,I'd like fluxbox as a option by default with Mepis menus and look and fell.

"I'd like fluxbox as a option by default with Mepis menus and look and fell."

Guys, can we please try to keep the posts relevant to the thread?
Thanks.

Ubuntu better than Mepis on hardware detection? Not with the two computers I have. I don't know enough about the rest of the discussion to opine.

Part of the confusion here may be in emelbert's four statements for the vote. I think the most fundamental question really is "Should Mepis proceed with the DCCA kernel, or fork (back) to a different one?" From the posts above of msandersen and TGBaker, it appears that DCCA, Ubuntu, and Debian are working collaboratively on a joint kernel. I believe Mepis definitely should use the DCCA/joint kernel. And all of us users should strongly support and appreciate the collaborative effort.

Bob Howard

I have tried out both Ubuntu and Kubuntu. I found both to be quite usable with the exception of sound.

Sound was flaky at best. It would turn off and on seemingly at random. This was a major complaint. I don't know if this has anything to do with the kernel or not. I did find hardware detection to be better. Even my old Advansys SCSI card worked with scanner and ZIP drive.

Even with some of the problems that have surfaced during the 3.4 release cycle, MEPIS is my Distro of choice by large margin. Warren has shown that he knows what he is doing, and I don't believe would go in a direction that would hurt MEPIS. I'm not overly worried about the politics. People complain about MEPIS having proprietary software in the OS Center, others complain about Ubuntu's practices. I'm for a great Distro that can keep getting better.

I voted yes - change if MEPIS would be improved, however, flaky sound would certainly not be an improvement.

My suggestion would be to release a testing version of MEPIS with the Ubuntu kernel to paying customers of the ftp site before putting it on the mirrors so those of us who are dedicated to seeing MEPIS improve could put it through its paces before a wider release.

CCJ

------------------------------------
What, now this is news to me, MEPIS Tools is what makes MEPIS better than Debian and is the main point of using MEPIS, if you don't like MEPIS Tools just use Debian. On my iMAC I use Debian, if MEPIS had a PPC Install CD, I would use MEPIS. I have never heard any one say a bad word about MEPIS Tools.

CCJ isn't complaining about the Mepis tools. (S)He is just staing what some others have commented.

There have been some people complaining about the Mepis OSCenter being closed-source. They think it should be released under the GPL as open-source.

But back on topic about the kernel:

I read a psot on the kanotix forum that argues:

"Ubuntu is taking for free every 6 months a (reduced to approx 15%) snapshot from Debian Sid, is changing these packages to make them entirely incompatible with the rest of the Debian world, and adds their own (also incompatible) backports later to this snapshot."

anticapitalista

Philosophers have interpreted the world in many ways; the point is to change it.

I'm certainly not complaining about MEPIS Tools. I was just pointing out that politics tend to come into play here, there, and everywhere.

Personally, the politics don't interest me. A strong, evolving, Distro that works does. If using the Ubuntu kernel keeps MEPIS working in this direction, use it.

CCJ (Him)

I decided to try Ubuntu and see what everyone was so excited about.
First I tried downloading the distro and tried the demo. It installed but didn't find my network card. No internet no dice!
Then I requested disks from Ubuntu which I recieved about a month later and tried the demo from that to see if there was a problem with my download; still no network recognition.
If it can't recognize a simple thing as a NIC then I won't use it.
Suse can find it, Xandros can find it, Gentoo can find it but not Ubuntu. Mepis finds it too and is my preferred distro, I hope it stays that way.

A great many of the questions posed in previous posts may be answered by the DCC Alliance 3.0 Release Notes at http://www.dccalliance.org/releasenotes.html and the notes should also reduce much of the apprehension about Debian package compatability.

I have used a lot of other distros before mepis, and mepis kernel simply does a better job of hardware detection.

I tried ubuntu when it first came out right before I tried mepis and I was not impressed, not to down another distro, just in my opinion it did not mesure up.

I think mepis is the best distro I have ever used, and I have tried a lot.

"If it ain't broke don't fix it"

I am sure with some improvement ubuntu kernel could be as good as mepis?

Thanks,

ahsiii

I think it's a great idea. I've been a loyal Mepis user since '04, but I must admit that when hardware detection became a problem for my wife (i.e. USB flash drive, firewire external harddrive) I switched one of our machines to Ubuntu, and it really is smooth. If that kind of hardware detection can be combined with the well-thought out Mepis distro, and the amazingly supportive community, the result can only be a winner.
Whatever the outcome, thanks Warren and thanks to all of you.

gitfiddler

I prefer Mepis largly due to good Modem and older P3 hardware support, which Ubuntu lacks, I hope we wouldn't loose that. Though, I do prefer Ubuntu's more professional and stylish desktop.

Example - Ubuntu Dapper Drake (version 6.04 -- a future release) recognizes Broadcom 43xx wireless natively

See on Linux Questions http://www.linuxquestions.org/questions/showthread.php?p=2083612#post2083612

I really do like Mepis. Ubuntu has many advantages, however. There are problems with bankrolling free software --- but there are problems everywhere, and as I read the Ubuntu forums and the Ubuntu laptop project more, I realize that money does buy results, even in the free software world. Many people truly like Ubuntu, and it has many good features.

Here's an example from Linux Questions. I'll post my response, which is also on mepislovers.

I'll send this separately to Warren and Dave Neil, because I think it's important they see the benefits of using all the new advances.

Mike

===
Broadcom native driver and Debian and DCC -- very important

It's now part of the kernel tree

At some time in the future, many distros will recognize Broadcom 43xx automagically, as they do ipw2200, with a kernel module that will be added automagically when the live CD routine or install routine sees the Broadcom 43xx.

The plan for the DCC kernel is that it will soon be very similar to the Ubuntu kernel, and so it should soon contain this driver natively.

I believe Dave Neil of Xandros and Warren Woodford of Mepis will soon agree and announce that they will use the Ubuntu kernel, or at least nearly all of it, as the DCC kernel or the basis for the DCC kernel.

I hope I'm not too premature in mentioning this here.

They recognize that some good has come from all of the Shuttleworth-Ubuntu testing of laptops, etc. and the expenditure of people and money in Ubuntu, and they want to retain as much of this as possible in the DCC kernel ...

This is the advantage of a native driver ... it will make Linux in general much more desktop and laptop-friendly. It is much closer to point-and-click for drivers. It should help in Linux adoption on the desktop.

I have not had problems with ndiswrapper either, but a friend who is an experienced Linux server admin using FC4 on a home system reported spending 4 hours and being frustrated trying to get a Broadcom 43xx to work with ndiswrapper. He was looking for a card that would work "out of the box" with FC4. I suggested that he keep trying, that I would be happy to help, but I haven't heard any more about it yet. There is also the problem of the hotplug/udev switchover -- so anything that makes hardware detection easier is welcome ...

Getting too long-winded ... but I think this sort of work is very important. I may try a new kernel with my laptop to see if I still need ndiswrapper.

Keep up the great work.

I will try to send part of this to Warren Woodford and Dave Neil so that they will see more proof that a DCC kernel needs to include all these sorts of advances, no matter where they come from.

Many of the advances, such as this one, are showing up first in Ubuntu because of Shuttleworth's public relations campaign and investment there.

Mike

1. There seems to be a lot of support for the idea that Warren is in the best position to judge - and I agree with that.

2. The objectives of the DCC seem very valid and MEPIS, as a founding member, should use the DCC kernel - if that is based on, or the same as Ubuntu, so be it.

If Mepis is going to remain a member of the DCC, and the DCC is going to use the Ubuntu kernel, then Mepis will have to do the same or leave.

I really don't care as long as it works and is good for the Linux community in general, and any collaboration is good for the community... normally.

The negative comments about Ubuntu usually have to do with personal preferences, and not actual facts. Ubuntu is not Mepis and is not intended to be. Use what works, and if they collaborate to produce something better, then I see that as a good thing.

I just removed ubuntu and kubuntu off of my machines and installed Mepis on all of them. as for better hardware support, if that was the case they would still be on the computers. Ilike Mepis because I don't have tospend days configuring everything. with Mepis I only have to install one package DOSBOX. DON'T CHANGE MAKE WHAT YOU GOT BETTER.

Here is another DCC post that may help understand better of what's goin' on. Ubuntu is helping with DCCA though not a member:

There's been quite a few emails about various DCC 3.1 related projects
in the last month or so. To clarify what we're doing here's a brief
summary of what is currently happening with DCC3.1 and its projects.

Joint Engineering efforts:
We're now working with Ubuntu in the following areas, kernel, Xorg, and
laptops. We're still in the organizational stage with these, but it
should move very quickly over the next week into a structured plan, and
actual hands on work.

Kernel:
We'll be moving to the 2.6.15 kernel for DCC3.1. Ben Collins is putting
together a proposal on how we can structure the development tree to
satisfy Ubuntu, Debian, and the DCCA. The idea goal would be to have ABI
compatibility. We'll post the proposal up on this list once we have
received it for everyones input. The joint Ubuntu-DCC-Debian kernel may
or may not be used in DCC3.1. It is very dependent on how well over the
next week we can organize to development structure. Xandros has been
working on a 2.6.15 kernel and we can use this if need be for DCC3.1.
We'll have a tree that Warren can build official DCC kernels from within
the next week or so. The kernel project is a very important project
with long term benefits, so we are making sure that we organize it in a
way that benefits everyone.

Xorg:
We're moving to Xorg 7.0. We'll have to work through any Debian Sarge
package dependency issues. Daniel Stone from Ubuntu has packages done of
Xorg 7.0. It is my understanding that the Debian X Strike force is also
using these as a basis for future Debian packages. That being so, we
should be able to all work together in the future to ensure
compatibility etc.

Laptops:
We're working with Matthew Garrett from Ubuntu. He's in charge of the
laptop project there. He has provided us with a outline of the tasks
he's currently targeting or working on. From that list we noted the key
binding task. This task involves mapping out of laptop "extra" keys.
(email, web,etc). Linspire, Ubuntu and Xandros are putting resources
into this project and will shortly begin work. If anyone else wishes to
help with the laptop project please let us know.

DCC3.1 Packages:
In approximately 10-14 days we'll be sending the kernel and Xorg
packages to Progeny in order for them to create the DCC3.1 archive.
From there other members will be able to begin testing with these
also. Our target DCC3.1 release date should be some time in mid to late
February if all goes well.

Thanks
Dave