This is the sort of question I daren't have asked at the sidux forums, but feel that I now can with regards to the more independent smxi.

I've been burned several times, once when the kanotix team split and when sidux did too. I've also seen other good and popular projects get dumped by devs that just didn't have the time.

H2's very cynical thoughts on open source development (http://techpatterns.com/forums/about1298-30.html, for example) also lead to asking this question.

So, the question is who would take over smxi development if h2 couldn't? I'd hate to switch to debian squeeze+smxi only to have to change again a year later.

I hope this doesn't cause offence, it really is something which has to be asked now for most distros and even major open source projects, the question of their longevity.
Back to top

smxi isn't required to run Debian Sid/Testing, it's just a light wrapper for various Debian tools and methods. In other words, it tries to de-geekify Debian, which is created by and for geeks, but is a nice base for your operating system, and apt/dpkg are excellent package management tools, far better imo than rpm or any other one I've tried, that's why Ubuntu uses Debian as a base, it's why Mepis, Mint, etc, use it.

One of the nicest compliments I got was from an oldtime user of pclinuxos, who said in his opinion, smxi was the best system management tool for real world end users since the pclinux tool. In other words, it made user unfriendly actions user friendly.

If I gave you a list of the core systems in Debian that are run by only one person, you'd reconsider using Debian almost immediately, and these are systems that actually matter to the end user. The openssl security fiasco that compromised all ubuntu and debian servers in the world for over a year was caused by one guy not understanding the code he decided to modify. And you know what? He's the same maintainer today. What does this mean? It means that if I need a locked down secure system, Debian is not in the running, nor will it ever be.

Every project is like this, almost, there's a handful at most of key devs on the larger ones, on the smaller ones its often just one person, for example Debian relies on basically one person, who is really tired of maintaining iceweasel, for iceweasel. That's why I don't use Debian's firefox redo but rely on the actual Firefox archived package direct from Mozilla, which always works, and more important, which gives me the new security releases the same day of release, or the next at the latest, automatically.

This is also by the way why it's probably best to use straight Debian than a derived debian distro, at least in my opinion. And it's why smxi always tries and tried to keep to reasonably safe alterations that would not negatively impact the long term stability of the system, I learned that from kanotix, that used mean ugly hacks to get things done, that made the systems almost unusable without expert modifications over time, in fact, I just found one last one while upgrading last night that blocked the upgrade for a while.

If you want software issued with a warranty of continued support, you have to pay a company for guaranteed support.

This is why companies pay Redhat or Ubuntu for support contracts, and it's why ALL GPL licenses say this:



Since smxi is just a wrapper, in no case would the system or user ever require it to run their system, it's merely a convenience to make things easier.

If someone appears who starts contributing code in an intelligent competent manner, then the various scripts I do would have a co-maintainer, but since that's never happened, just as it's really unusual for most other smaller free software projects to get competent contributors who understand the code base, things will just roll along as they do.

I use a lot of stuff done by one guy, or small group, like Aqualung for example. But I know they might get tired and move on, and I also know that the odds of competent replacements appearing are not very high. But that doesn't matter, I don't care, that's how using free software is. Overall I find that most projects do ok long term, some don't.

For example, I really relied on the high quality kde 3.5 desktop, and now that kde 4.4 is in Debian, and it's still just about as unusable as Windows Vista, if not worse, I'm really kind of screwed, because I hate kde4, don't really like Gnome, and find the lighter desktops simply too simplistic or bug filled to use for serious work.

Same goes when a new kernel breaks my hardware, because the devs changed the kernel APIs, again, as they do about every 6 months.

These are all part of what is meant there is no warranty of any sort with free software, that's how it is, it's best to simply accept this reality and not worry about the future beyond keeping up with changes you'll have to spend your time (nothing after all is actually free) to learn how to deal with. I, now, for example, have to waste time learning how to do grub 2 configurations.

Remember, when you pay nothing, you should expect nothing, except in hopes that the goods and services you are receiving for free actually continue to be offered for free. The GPL guarantees that the code will remain free and available, but it cannot guarantee that unpaid developers continue to do the work for free at often great personal cost.

Personally, I worry more about group think than individuals. Sidux users were happily told to use smxi, they loved it, then one day they woke up to find that all support had been withdrawn, mention of the product, which had been in sidux officially via the installer, was suddenly grounds for bans from their IRC, etc. So that's what I worry about, the kde group made some horrible decisions with their new desktop, grub2 totally destroyed usability of grub for anyone other than a total geek, but you know what? - see the no warranty quote again. That's what using free software means, and it's why companies pay Redhat and other vendors a lot of money for professional level support for the packages the vendors support.

I'm not cynical, I just read the license text and I watch things break constantly, undergo usability failures caused by incompetent or inexperienced or simply bad programmer or programming groups taking advantage of the no warranty protection free software guarantees all users. The myth of the endless pool of skilled programmers waiting to take over every project out there should the lead dev falter is just that, a myth.

This is why Apple bought Cups print system, for example. It's why mysql is owned by a corporation. They needed that warranty of usefulness, and the only way to get it was to buy the project. It's also why the only usable office suite was paid for by a corporation. And it's why companies employ linux developers, to make sure what they need is done. And it's why redhat will pay for things they need but aren't getting done, like Liberation fonts.

Not to go on, but you have to keep front and center that the reason you use free software is to be using free software, not to have a security blanket. Microsoft does great work with long term support, better than anyone else, XP still has full support almost, and it's 9 years old. Redhat I think will offer up to 5 years, and have just added an expensive extension program so you can keep getting that support for extra years. But this costs money, just as in the case where I don't have time to do this work, and nobody is helping, which almost nobody is, then I will not sacrifice my life for this project, but I will make an effort to keep the nice parts rolling along as long as people find them useful.
Back to top

techAdmin/h2, thanks for sticking with smxi, despite it being an thankless chore sometimes, it really does make life easier for users of all types, especially with kernel updates and proprietary video card drivers.

It is easy for relative newbies, but it also allows people more understanding of what is actually going on rather than some mystery script, click, click, terminal pops up disappears and stuff is broken with no clue as to what happened.

I agree with your summation of the Desktop environment blues.
I have one laptop still running KDE 3.5.10 and it really makes me wonder why they changed, but I guess it was becoming an unwieldy mess under the hood and there was pressure to make it look "new and improved".

That said, I find that slowly KDE 4 is improving from unusable to functional with annoyances, although part of that might be my just getting used to it.

I think with the Testing freeze things are stuck at 4.4.5.

If you feel like taking a chance you can grab 4.5.1 here:
deb qt-kde.debian.net/debian/ experimental-snapshots main

Works OK on this box thus far, I may try it on my laptop too.

Usual disclaimer: Experimental! Don't try this at home kids, don't stick it on your company's server.

Marcus
Back to top

Well, you can still have KDE 3.5.x in Squeeze (and hopefully beyond) with Project Trinity. The version they're at now is 3.5.12.
Anyway, I'm really grateful for your scripts, they really make managing Debian a breeze.
Back to top

sickie I might check that out, I really cannot see anything at all that kde 4.4 does that makes my life nice, nothing.

At least 3.5 works, and if you can use kde 4x apps on the 3.5 desktop that might work out fine, as it is now, I can't install any 4.x apps because it will break the 3.5 system.
Back to top