MEPIS Talks 4
From MEPIS Documentation Wiki
Note: this podcast was created 8/13/2007, and you can download the original audio file here.
Note: This is not a word for word transcript but typed to get the points across, please feel free to go through and correct 16/09/2007.
Matt: Hello everybody, welcome to MEPIS talks, this is episode 4 I believe and Warren and I are back from a long hiatus. We haven't done one of these in quite a while and trying to get back into the groove of doing them. We've been pretty busy lately. I think you're still there, Warren?
Warren: Yep, I'm here, hi everybody. One of the things that happened is that Matt and Aaron moved and we just got off track with doing it but we're back and I hope we'll be doing this, some kind of a pod cast, every week now.
Matt: Aaron and I got settled up here and just trying to get everything settled and finally got into some kind of a groove, a schedule, and figuring things out and get back into the groove of doing these too. So, welcome, everybody again. One of the things I think we have ironed out too, is our voice quality should be better this time. I know Warren's wasn't too bad but mine was pretty bad but we went ahead and looked at some different software and took a look at some different things. We tried using skype before and that wasn't working. So, any suggestions on that one?
Warren: Yeah, what we're using now is Gizmo which is a different VOIP application and the nice thing about it is that the audio recording capability is built right into it and as far as I know it's the only one like that. We both looked at some different applications and with all the others it was necessary to use a separate application for doing the recording, like... and it was really difficult getting it all working together but in this case that's just taken care of with the audio recording capability built into the app. So here we are and ready to go, so what are we gonna talk about?
Matt: Well, the first thing I think we need to talk about since we haven't been talking for quite a while, to the community is switching back to Debian. I think that's one of the big things that hit the news lately.
Warren: Yeah, it has hit the news now and it might even be considered old news. But for anybody who hasn't heard the full story, we were having problems just getting everything the way we wanted and we were really wanting MEPIS to just be upgradeable so that people didn't have to do reinstallations every 6 months or a year or whatever and we just really couldn't get a handle on that with the codebase that we were using although we did a lot of our own upgrading for 6.5 and I think we did a good job with that one to accomplish as much as we did but it just didn't look like the direction we were going was going to work so we decided to go back to using Debian but not using it the way it was used before. MEPIS Simply is being built on a Debian Etch base but we're going to use that to use a really solid, stable foundation for 7.0 but then we're going to use the very best, latest packages from upstream to add on the latest and greatest user applications on top of that stable core so hopefully, what it looks like the result is going to be, is that we have a really stable build, we have maybe 15'000 extra applications that people can install from the Debian pool but at the same time we're gonna have a core of a couple of thousand maybe even 2'500 packages that we are maintaining making sure that they are up-to-date and that all of the latest and greatest, Firefox, Open Office and KDE and so forth, are available on top of that stable core and with that, we figure, we're gonna have the best of both worlds: stability and the latest and greatest stuff you wanna use. And in addition by taking this approach that we will be able to continue to update and upgrade these packages so that going into the future for the next couple of years, you won't have to do a reinstallation, you'll just have to do an upgrade and from time to time, maybe point to another pool that has newer stuff in it.
Matt: right, so, what is the roadmap for MEPIS 7? What are we looking at here timewise for getting it out the door and what needs to be done yet?
Warren: What I sort of implied was that the goal for 7 is not so much to add new stuff as it is to have the latest and greatest user applications and to switch over to this Debian base in such a way that we will be able to upgrade into the future so the focus is to get it stable and get the latest stuff in there and make sure we're upgradeable. The timeline for that is that I'm expecting we'll have Beta 1 this and a total of 4 betas and a couple of RC and if all goes well, we'll be final by September. From there, we will, as new apps come out we'll build those new versions and if it's a security fix or minor update say from Firefox 2006 to 2007 that will probably just be available as an update to the current version but there will be some point where we will have enough changes that will warrant having a new pool that will represent a minor update to MEPIS and at that point we'll do a release associated with that. All along the way, though, people who are already running MEPIS will be able to update to the new versions from the available pools. They may not want to go to the pool with the experimental and unstable stuff we're still testing; they may want to hold off on that or they may want to do it, depending on how close to the edge they want to be. People who want to be really stable can wait for the release, then add new stuff. People who want to be on the edge can go with the stuff early but in both cases they can do it by adding packages rather than a new installation.
Matt: That's actually a pretty big thing to be able to do that. How long do we plan between upgrade cycles? Say we come out with 7.0, how long will we be able to do updates before a new install?
Warren: This requires a lot of work and skill. In a way, we are going into new territory because I'm not sure what other branches of the Linux tree does or doesn't do but I think in the Debian branch we're doing something that has never been done before. In fact, this isn't all my idea, this is a variation on the idea of the Debian common core. The consortium that was proposed a couple of years ago by Ian Murdoch, that group of about 10 entities, quite frankly couldn't pull it together because I believe, of the different interests of the different parties. It wasn't just a matter of technically solving the problem, there was a lot of politics involved but the idea, I feel, was a sound idea of having a solid, stable core and then adding the good stuff on top of it. As far as I know, we're the first ones going to be doing this in the Debian world and there's a lot of uncharted territory there. But if all goes well, we'll be putting all of our focus on this build until the time approaches for the next major release of Debian. When Debian Lenny is approaching the Debian freeze, which will probably happen, I don't know, a year and a half from now. At that time, we'll start looking very seriously at a new major release based on Debian Lenny. At that point, it isn't clear, whether we'll be able to provide an incremental upgrade for the people who have the current version of MEPIS or not. Maybe we will. We'll learn a lot in that year and a half and maybe what we learn will make it possible. At the moment there are just too many unknowns to say for sure. Probably, I'd say a year and a half and maybe even then an upgrade will be possible.
Matt: Well, that's great. Anything else about MEPIS 7.0 and what's going to make it different, besides the upgradability bit, that's huge, but is there anything else you want to inform the community about?
Warren: Well, I'm not sure. I mean, the upgradability is the big thing. 7.0 does have a new kernel but that's to be expected. In this case, you know I gravitate towards using a stable kernel, not an experimental one. I've said before, that there's a pattern to the way that kernels are developed. The kernel team goes through a cycle, led by Linus, obviously, where new technologies are introduced all the time and sometimes those technologies destabalise the kernel for a build or 2 or 3. You can pretty well predict that once a year there will be a really stable kernel. I thought that maybe we'll go with the 2.6.20 kernel and then I looked at the 2.6.21 and it looked like it was pretty close to the 20 but with some improvements. I decided to look at the 2.6.22 kernel and the fact is that it looks really good. It looks like a really stable kernel and based on that it looks like we'll be going with the 2.6.22 and that will give us some really up to date hardware compatibility particularly regarding the latest Intel chipsets and things like that.
Matt: There are some things in the kernel that may affect some things when people are doing an install for the first time, once we get into the beta cycle and one of those is that your hard drive names may have changed since 6.5. Before, your hard drive may have been named hda and now it'll be named sda, is that correct?
Warren: Yeah, that's pretty standard now. One of the things that really destabilised the kernel after 2.6.16 was the development of a whole new generation of what are called block device drivers which are the drivers that handle accessing hard drives and CD ROM drive etc. And starting with 2.6.17 the kernel became very unstable and this is why MEPIS didn't upgrade kernels for a long time. The kernel is going through some real growing pains in this area and the whole new generation of drivers has shifted in favour of naming devices sd-whatever rather than hd-whatever regardless of whether they are USB-based or SCSI-based or IDE-based. No matter which they are all called sd- with these new drivers. There's another thing that can come into play related to that. For example with my Mac Pro, which is my primary development machine, the hard drives, the second and third drive letters were reversed due to some oddity in the older driver that is now fixed in the newer drivers. What used to be sdc is now sdb and vice versa.
Matt: Also, the kernel is now being built SMP by default.
Warren: Yeah, this is to some degree, you might say, an experiment. There are a number versions of linux out there that simply offer an SMP kernel and that's it. Initially, MEPIS offered a single processor kernel and the multi-processor kernel was an option that you could install if you needed it. I felt more comfortable with the single processor kernel. It's a lower common denominator and is more likely to work everywhere, or at least it used to be. These days, the kernel is supposed to be very robust if you have a multi processor kernel and use it with a single processor machine. The interal code in the kernel is supposed to be very good at handling that. I'm hoping that we can go with the multi processor kernel as our primary kernel and that that'll just work for everybody. It may turn out that we need a single processor kernel for some people. If that doesn't work out during the beta period we may have to swap back to using a single processor kernel as a default. But these days a lot of the machines that are available are multi-processor machines. For example, if you have a newer machine that has an Intel processor that calls itself a Core Duo, that means that there are two cores in the processor. If it says 'Core II' instead of just 'Core' it just means it's a new generation. That's so confusing that they have 2 and duo, one means that they have two processors and the other is just newer. If it's Duo, you have two cores. If it's II, it just means you have a newer piece of silicon.
Matt: Even the old Pentium 4's that had hyperthreading, that also is considered multi-cored.
Warren: Yeah, it's not the same thing but it can be taken advantage of in the kernel and performs better if you have SMP or multi-processor kernel. While we're talking about that kind of distinction, there's another thing we can throw in to educate people: is it 32 or 64 bit. Where it's very clear that it's AMD 64 bits and obviously MEPIS is in both 32 and 64 bit versions. If it's an AMD processor it'll say 64 if it's a 64 bit processor or if it's X2 it'll be 64 bit with dual processors but with Intel even though it's called AMD64, some Intel chips also support that. All of the core II's are also AMD64 compatible with what they call EM64T, it's the same thing. So you can run 64 bit MEPIS if you have a core II or if you have a Intel Pentium D, which was also EM64T. I know that this can get confusing for people when they're trying to choose which version of MEPIS to get. The safe version is 32 bit but if you have a 64 bit processor you'll get a little bit better performance if you get the 64 bit version.
Matt: The rule of thumb is, if you're unsure, just get the 32 bit version as it can run on both 32 bit and 64 bit processors whereas if you get the 64 it will only run on 64 bit processors so only get that if you're absolutely sure that you know what you're getting into.
Matt: Some of the other things that are coming down the pipe are KDE4. What's MEPIS been doing with KDE4?
Warren: Well, this past week I put out iso for 32 and 64 bit. I did some testing, some experimenting with KDE4 which is currently in a beta 1 state, where there is a lot of work to be done before they get final, it'll probably get final in October but might be a little longer, but I did this work with KDE4 last week because in the future, obviously with our new plans, KDE4 will come along and at some point we will want to offer it as an upgrade in the Simply Does series and in order to do that I need to make sure that MEPIS is ready for KDE4, that we aren't going to run into any stumbling blocks or serious problems when KDE4 is final. So, I took the experimental packages from Debian SDN and from Ubuntu universe pools and tried them out. I found that the Ubuntu packages worked a little better this time but that they compile (they work with the 7.0 base). I had to add a few updated packages and new packages to MEPIS to support things that are going to be in KDE4 but this in no way interfered with what's already in Etch. So it's following the vision, the plan, that for us to be able to support KDE4 it looks like we'll be able to just add stuff on top of Etch, we won't have to break Etch in any way. We'll just be able to put this stuff on top and we'll be able to offer KDE4 and after I got that work done I thought, hey, this is not something I would normally release to anybody but that I thought people might want to have a look at it. At least you can probably start up KDE4 and see a little bit of what it looks like. It's clunky right now, it's sluggish. It requires a relatively powerful machine at the moment. Older hardware, slower hardware, underpowered hardware it probably won't run with at the moment because it's still being improved but I thought I'd put it out there as people may want to take a look at what we're doing and maybe they want to take a look at KDE4. So it's there (in the testing pools), anybody who tries that out really needs to keep in mind that this is a really early thing. They shouldn't be critical of it. There's a lot of people working on KDE 4 and I'm sure it'll be great when it ships and you can see right now that there's some good things coming along with it and it's really gonna be interesting and also nobody should be critical of the Debian or Ubuntu developers. They're doing a monumental job, this is a huge amount of work they're doing and everybody should be really thankful and appreciative that they're doing this work to bring this code into the Debian branch of linux and I wish them well. I for one, really appreciate what they're doing.
Matt: Great, so KDE 4 another great thing to look forward to coming down the pike.
Warren: Well, let me say that for us, KDE 4 will absolutely not be ready in time for the 7.0 release and it being a really major release of KDE, it'll take a little time to get all the kinks out of it but we'll have packages available. My plan is that every few weeks, I'll build it again. Every time that there's a significant improvement from upstream, I'll do a rebuild. At some point it'll be put into a MEPIS experimental pool which does not yet exist. In order for us to do what we're going to be doing, not only will we have a stable release pool but at the very least have an experimental pool and an unstable pool. I don't know whether we'll have another level of testing pool or not but we'll be having more levels than before so people can choose how close to the bleeding edge they want to be.
Matt: Right. Anything else you want to add about any of this? We have one announcement left after this. If you have anything else to add to the podcast, or we'll get on to the announcement or question really.
Warren: Well, I can't think of anything at the moment, I'll probably think of something later.
Matt: Well, we have a quasi announcement. We're looking for people to ask Warren questions and to actually come onto the podcast and be a part of it. Maybe even be a guest host one night if I'm not able to find the time to do it. Be a guest host, ask Warren any questions about MEPIS. Go ahead Warren, add to that please.
Warren: I think that this is an opportunity for people. I mean, I'd like to be able to talk to people all the time. What happens is people send me emails but I spend all my time, practically, sitting at the computer typing and I get tired doing that and I'm not really thrilled with answering emails for hours and hours and it seems that having people talk on the podcast, or being a host and asking questions is a good way to give, you the listeners the opportunity to be heard to say, what you want to say, to ask how you want to ask. Now if you want to come on and tell us how horrible MEPIS is and how much you like some other distro, we probably won't want you or at least won't publish your podcast but if you want to come on and ask some real on-point questions or just want to ask about stuff related to MEPIS or me or linux or whatever, you're welcome to. The general idea is that this is an opportunity if you want to take advantage of it and Matt's going to set something up so that you can let us know.
Matt: What we'll do is create a form that anybody who wants to participate in this to go ahead and fill out the form and I'll put a link to the form in the show notes when we post the podcast so you can just go ahead and look in the show notes, it'll be at the top of the page. Just click the link and fill the form and we'll go ahead and pick somebody. We'll contact them and let them know when we'll do the podcast and include them in what we're talking about.
Matt: Even though we have a lot going on I really don't have a lot else to say at this point in time. I think we'll wrap things up. We'll encourage you to fill out that form link if you want to chat to Warren and come visit us. If you like what you're hearing go to the MEPIS store and give a donation, sign up for a subscription, anything like that is a big help. I'm going to say goodnight to everybody or goodbye for this week depending on when you're listening to the podcast. I'll see you guys hopefully in another week or so and hopefully we'll be doing these on a much more regular basis.
Warren: Yeah, goodbye everybody and thanks for listening.
Matt: Alright, see everybody later.