Sidux

"Is sidux a Live-CD system?"

This question seems to be absurd at first since sidux *IS* a Live-CD system. But is it really?

To me, a Live-CD system is a system that I can take to anywhere and have *my preferred* tools/environment ready whenever I wish to. If the purpose of a system is just to let people browse the Internet, do some online chatting, and/or play some music/movies, why bother reboot to a slow system that is on CD/USB, whereas the system already on HD is much faster, and it already has those tools that the system is built for? I know the difference is mostly Windoze and Linux, but I don't care. I only want to get the job done, not to be a Puritans that do not do certain things because of religious believes. If I have to accept what is force on me, if I can't have a system that works my way, I'd rather settle with whatever it is there on the existing HD, instead of spending time/energy/money on downloading/burning/testing/getting accustomed to another system and waist time to reboot to a effectively slower system that imposes more restrictions. One example, out of box, all my changes to the system will be lost after reboot, at least until now, and I have to do the very same thing again and again. Moreover, mp3/movie playing feature is not possible out of box.

To me, Linux is all about freedom, freedom to dream and freedom to break away from any restrains imposed on me. If you also believe so, and wish sidux gives you such freedom to have your own preferred live tools/environment, read on closely what's excerpted below.

PS. personally, I will only use sidux to install Debian on to HDD, not as a live system. But that'd only be my second choice, because grml gives me more freedom when it comes to customization. Moreover, the installed grml system is a near live system — it will still work if I tar up the whole system and untar into a brand new system with entirely different hardware. These are the unbeatable advantages grml over sidux as of 2007-09-02.

PPS. IMVHO, I think sidux needs to straight out its mission statement as the very first thing, i.e., the strategic planning is very important for such community driven project. Due to the unstable nature of SID, the most appropriate target users of sidux is experienced Debian users, who know how to fix things when the system breaks. But for those users, a customizable system that works their way is extremely important. This is an issue that never get properly addressed by the sidux development team. Currently sidux seems to be targeting and attracting Linux newbies, ie, people who have no experience with Linux at all. However, an attitude of "accept it, don't touch it" will only make sidux a toy rather than a tool. I mean, after the intimidated Linux-newcomer start to have some confident about Linux and want to "touch it", ie, make it works their way (software installation being a prominent one), things will most likely to break. IMHO, this is no good to the Linux community as a whole, because they will be intimidated and frustrated and blame *Linux* to be too easy to break, not knowing that they've made the wrong choice at the first place.

documented on: 2007-09-02

documented on: 2007-08-12

http://sidux.com/PNphpBB2-viewtopic-t-4083.html

> I was making an experiment, remastering the ISO of Tartaros after adding
> some stuff inside...

Forget "ace-something" - sidux is not mastered in such trivial way. You might want to follow the user based discussion about re-mastering here: http://sidux.com/index.php?module=pnWikka&tag=LiveCDRemastering. However, the sidux dev team strongly recommends to stay away from re-mastering.

The better and proper way is doing your own master - make yourself familiar with the official sidux mastering scripts in our SVN.

Jun 05, 2007, slam

http://sidux.com/PNphpBB2-viewtopic-t-4538.html http://sidux.com/PNphpBB2-printview-t-4538-start-30.html

> > 3) Optimize file allocation on CD so that CD/DVD drive spends less time
> > seeking and more time reading contiguously (maybe it would be necessary to
> > 'preread' some file during a boot sequence)

> > ... I have recently discovered that sidux lacks smartctl utility ...
> >
> > Please, include smartmontools as well.
>
> Quote slh:
> I demand the bikeshed to be green.

And moved 6 inches to the left. Laughing

Posted: Jun 30, 2007, Scrooge

birdie, in case the above comments are too obscure, I'll answer:

> I assume this request is about as silly as it sounds. ...

Not so silly, but simply requesting that we order the entire contents sequentially from "debian/sidux chroot" -> "squashfs" is easy. Writing the code to do it in an automated manner fit for our build scripts is not.

Therefore, much can be said about it, but without concept or code on how to achieve that, words are futile.

kelmo

Knoppix is built in a very different way sidux is. Actually it is always a manualy adapted and upgraded re-master of itself (By the way, Kanotix, too). If you manually (re-)master your build it is quite easy to do some file sorting (and bad hacks). As Knoppix comes as a huge packed DVD which is usually (and should be only!) used as a live CD there might be some reason to sort files on the disc.

sidux devs at the other side have created sophisticated build scripts which run fully automated and use the pure material from Debian Sid. This makes sidux 100% Debian compatible, and no hack will later on break your system with an upgrade. sidux is optimized to to be installed to HDD, the live functionality is well working but not the main focus.

With Tartaros I did some tests with interrupting the build process and do some manual file sorting. The resulting speed improvements at boot time where less than 2 seconds - so there is no reason to go further with this as sidux already boots extremely fast compared to other distros.

Posted: Jul 01, 2007, slam

All this discussion suggests me that, should a sidux remaster tool be available, all the requests would cease, as people would be too busy to make it "their own way" Laughing

I did it manually some time ago, but an automated tool would be awesome!

> I did it manually some time ago, but an automated tool would be awesome!

That exists from beginning, you can use the original sidux build scripts. http://svn.berlios.de/wsvn/fullstory/fll-builder/trunk/?rev=0&sc=0

I wrote a tutorial in german (the bold warnings and disclaimer are important): http://sidux.com/index.php?module=pnWikka&tag=LiveCDMastern

Blue Shadow - Jul 01, 2007

	The tutorial is deleted as of 2007-08-12.

> There's no specific need for a gui, in fact a text UI could be enough.

I think even a text UI is needless, because the build script is very easy to use and you should know what you are doing if you want to make a custom cd.

All that is needed for common use is a certain degree of stability (before that, the release dates are the best time to use the script) and a manpage.

But as slam said a few weeks ago in the german board, a gui would be nice, if somebody wants to code one.

Blue Shadow

Nope. If that happens i will never take responsibility for making changes that stuff up a user interface, nor do I plan to make an user interface.

Debian-Live project is planning these kind of things however.

kelmo

> in the future, this could be a *great* tool, and could give sidux a
> "make-your-own-sidux" bonus which is something very special.

i had a butchers at dreamlinux the other day, that has a good, and easy remaster tool, works in a few clicks- very impressive.

wegface - Jul 08, 2007

For installation purposes sidux defaults are excellent as they are, we can apt-get any extras we need and tweak user settings later.

However for those of us who travel a lot, use different PC's but don't like most other OS's, a livecd with a few user-specific apps and settings is a very useful tool. Individual needs can never be met by the best defaults possible.

That script has been made available to download and a wiki page exists so it's certain to attract interest. But it's obviously not meant for general use (yet) because

cp: error: /boot/grub not a directory

for me too and no man page so I did the "unclean remaster" just for a handful of needed extras and settings, fortunately it worked. I'm not interested in dreamlinux, pclinuxos or any other distro, I like sidux and proper debian.

I can't see the point of a master/remaster for other than a personal livecd nor a gui (those who can't run a script or have some reasonable Linux understanding are probably not up for that job at all), nor why anyone should expect official support.

But maybe this could be discussed more openly in the "anything goes" department. I wanted to write about this before but got the impression it's a "taboo" subject from the few posts in the German forum?

dzz - Jul 08, 2007

As one of the primary authors of the sidux tool I'll say it simply:

The tool that makes sidux is just that, a tool that makes sidux. It is not intended to make private copies of sidux, its intention is to be a tool that facilitates group collaboration to build a targeted set of products that we call "sidux".

There will be no support for anyone who does not want to achieve this goal, which means no support for anyone whose primary goal is to build a personalised "sidux".

Reason: simply don't have the time, nor am I personally interested or bothered to offer such a service. Others (like dreamlinux, debian-live) are.

kelmo - Jul 08, 2007

I agree with kelmo's statement. I'd like to add this: I trust a distinction will be made between free(GPLv2/3) and supported.

The license the script(s) is(are) placed under encourages free distribution. But please don't be as bold as to ask support for a script, which you don't even understand; and which is used to create a distro which you intend to take from(remaster) but not necessarily give back to. That's where the support line in the sand is drawn, and I think that is fair.

locsmif - Jul 09, 2007

Personally I believe these two things have nothing against each other. Plus, such a tool would only add competitive advantage to the distro.

Btw: if the point is just about "how time consuming" the project could be, an adaptation of tools created for other distros to the existing script could maybe cut times dramatically. I'm not doing it myself, just because I totally lack programming knowledge, but maybe somebody else could capitalize on the suggestion.

jaegermeister - Jul 09, 2007

As far as i am aware at least, even back in kanotix days, anyone asking for help to remaster, has been given no support by any devs. Its just not on any of their wishlists. As its them with the programming knowledge, this isnt a feature thats gonna happen any time soon. Reasons for this (i think) are to stop a remaster which includes tons of non-free, maybe non-legal software being released by someone in a country where these laws are not valid. This could then directly compete with sidux, while doing nothing but stealing sidux's hard work. If its possible to remaster with a few clicks, or even a simple script, whats to stop someone or ANYONE doing this? Plus forums would be full of questions from people using non standard, non supported fake "sidux" installs. It opens many cans of worms.

N.B: this is just my 2 cents. This opinion may be completely different to sidux devs views.

wegface - Jul 09, 2007

This is in fact a true motivation for opposing the process. Politics and not technical reasons.

Then, the next step of analysis, would be to examine whether such effects are actually happening to those distros who sport such tools or not.

jaegermeister - Jul 09, 2007

sidux won't stop anything, compared to your prior experiences everything we use to build sidux is out in the open and licenced under FOSS licences (even though we often do not advertise this fact for the reasons noted below); however these tools can be dangerous in unexperienced hands and are under heavy development with major and unannounced changes several times a day (a single typo might lead to complete data loss of your host system). I'd like to stress clauses 11 and 12 of the GPL v2 (excuse me for the all-caps text, but that is copied verbatim from the GPL v2, as published by the FSF - emphasis mine): http://www.gnu.org/licenses/old-licenses/gpl-2.0.html wrote:

Creating an ISO image does require some basic knowledge about and good experience with UN*X and the SysV Bootprocess, no pretty GUI frontend can alleviate that. There are already more than enough badly crafted linux distributions around and we have neither the time nor the motivation to support customised "sidux" derived variants, besides that knowledge going into forks is usually lost for improving the base system. If you can cope with the existing framework, have your fun, but don't ask us for illustrated step-by-step guides or support for your custom creations. If you intend to distribute your customised creations, honor the licences (which are usually GPL2 compatible) and communicate to your potential users to contact you instead of the sidux project for support.

My personal target audience for fll-builder are not end users but team members and as such it will stay published, but unreleased code; if you want to change that, start coding to make it failsafe before even thinking about a GUI - my priorities are elsewhere.

The entry level to contribute code and patches to sidux itself isn't that high… Basically it boils down to "show us the code and convince us why it's cool." You don't need to be a specialist - if you prefer it, you can also start to help us with janitorial tasks (take a look at our status reports for pending tasks, currently it is standardizing tools on ssft, gettext support and maintenance for bluetooth, gprs/ umts and modem/ isdn support) and learn about it while doing.

slh - Jul 09, 2007

> This is in fact a true motivation for opposing the process.
> Politics and not technical reasons.

Please retain some respect. I find your statement to be offensive.

kelmo - Jul 09, 2007

Yes, I have to agree, the idea that some user, who isn't doing the work, isn't spending their free volunteer hours, can say that's because x or y factor, is absurd.

jaegermeister, the next step of the analysis is very simple: do you want to do the work to scratch your itch? That's what's simple, it's hard to be more clear than kelmo or slh are being.

Here's how it works in case it's unclear to you: you volunteer some of the finite amount of creative free time you have in life to do what you envision, as a developer. Maybe something isn't working, and it annoys you, so you fix it.

Maybe you create something you'd like to see exist because, well, because you'd like to see it exist.

So if a proposal or idea fails to meet this standard, that means nobody is going to be interested in spending any more of their free volunteer time on the idea. So at that point, the person who thought it was a good idea has 2 options: forget about it, or learn how to do what's needed themselves.

No one is obligated to do something like this thread is talking about, creating a fool proof mastering thing, that takes a lot of work and effort, and nobody gains at all. If someone feels that there is a gain, then they should start doing the work, since apparently this feeling isn't shared by developers.

Personally, while the latest work on the new livecd is definitely impressive, to me the point, in the end, is installing thing thing and using it. Live stuff is nice, it's cool, and it can be useful, but I don't think any project, except maybe knoppix, can really focus on that single thing long term and survive.

You need a large installed user base as far as I'm concerned, without that, there's no commitment to the community except in rare situations like cleary's use of livecd technology at his work, but then again, cleary puts a lot of testing and development time to get the stuff working in the first place.

I think you'll find that if you start programming, you'll soon see just why sometimes it's a lot easier writing a program for a single application than to try to make it stable enough for average users. For example, I'd estimate I need at least 10 times more code to make sm / sgfxi as fool proof and error resistant as I can, it's not trivial to do that, it's a huge amount of work, so unless you feel like doing that work, well, consider it a suggestion with no meaning since the desire for that functionality is not shared by others.

h2 - Jul 09, 2007

h2, I'd be totally happy to help the project by programming, the point is that I do not even know where to start from. The max I can do are trivial bash scripts to retrieve stuff from internet and start some make or other application.

This considered, I'm contributing to the community with bug reports, proposals and troubleshooting within this forum. So, in my spare time, I'm doing what I can, providing user contribution.

Then again, the issue of a customization tool is not a religion…. I have partially remastered tartaros myself without fll-builder (which I still haven't tried) following the online instructions. Instead, it's more about reflecting over a trend:

lately more and more distros are trying to give the user an advanced customizable experience, some through remaster ability (e.g. dreamlinux), others through an accurate choice of packages during install (http://www.go2linux.org/ultimate-ubuntu-edition-cd-1.4-gnarly-gnome). Not ignoring this trend would surely contribute in enlarging sidux user base. Linux is about freedom and customization, and today's users are even asking for further degrees of freedom.

jaegermeister - Jul 09, 2007

Tutorials - very simple: http://linux.org.mt/article/terminal - simple: http://linuxcommand.org - advanced: http://www.tldp.org/LDP/abs/html

raft/rtfs, you'll have all the freedom you desire.

locsmif - Jul 09, 2007

Personally, I'd like to see sidux enlarge its user base by having more users run it as their primary operating system, installed. That's what I care about, that's what will make a difference long term re helping break the MS type monopolies that are holding back things.

Live stuff is nice for users, but for me it's never been an emphasis, I like having the livecd, and since I can use it to install anything easily, that's fine with me, but the important thing is to get people running any free desktop, for real, installed. Nothing else will achieve the result of cracking the MS / proprietary desktop near monopoly.

Users who really run sidux installed are far more likely I think as a group to contribute help and support long term, since it's their actual system they are running, not a relatively static livecd thing. One reason I really like what kelmo and everyone have done with the new livecd is that it's very much like installed now, so it's easy for users to get a feel for it by actually running it, at which point hopefully they will proceed and install the thing for real.

If you're thinking in terms of expansion, then don't think small, livecd users are a tiny fraction of installed users, so installed is the logical place to focus as far as I'm concerned, that's what I do anyway. Having maintainable livecd stuff is important long term, it's the way sidux will continue to exist, but it goes no further than that, my point is that it's a LOT harder to make a script that works for a small group of focused people work for a large group of relatively unskilled people, and sidux doesn't gain anything by doing that.

There are millions of computers out there waiting to run free desktops, ideally debian, why worry about a tiny number of specialized users?

h2 - Jul 09, 2007