DSL-N :: Move from DSL-N rc4 to DSL-N 1.0

Re: Difficulty -- My current peeve is that Ubuntu 3dfx drivers broke badly during Edgy and they don't seem to be interested in fixing it, even in Feisty. One of my two geriatric boxes (750 MHz PIII, 950 MHz Duron) is 3dfx. Bug is documented, patch available & works. No fix in distro. All distros have issues, even DSL... ;-)

Re: DSL/DSL-N filling niche for tiny distro -- I agree. We disagree about means to enhance the base. Frugal is great, in its place. Traditional Debian has advantages of its own, including learning curve and easier to add current packages with Synaptic.

Small base distro, two supported install models, two different repository systems. Choices are good, IMHO.

Please note: 36% of DSL users use "Traditional Hard Disk Drive install," per DSL User Survey Results. Only 29% use "Frugal on Hard Disk Drive."

Re: Ease -- DSL/DSL-N are different from standard Linux. There are advantages, but there is also a learning curve. One thing that is true -- many DSL packages are downlevel.

I really think that it is easy for geeks like us to forget how much more we know than the average user. Again, the DWW reader comments indicate that even those Linux enthusiasts are much, much lazier than thee and me. Highly recommended reading.

When many of these people are chosing Mint over Ubuntu because Flash, codecs, multimedia, etc are pre-configured, you have an existence proof of Eric Raymond's "World Domination" points! For these people, Automatix is too much work -- pkgsrc is not an option!

I don't want to take away the things that make DSL/DSL-N what they are. However, I think that there would be many more users if it was radically easier to add desired current packages, even if these packages don't particularly appeal to the majority of current DSL-N users. (If they want to "waste" their memory on "bloat" -- more power to them! Personally, I'm happy two "waste" the memory for K3B.)

I'd submit that Debian Stable is a bit too stable for most folks. YMMV. The genius of Ubuntu is popularizing Debian. They aren't perfect, but they seem to have done a pretty good job balancing the risks/benefits of being a bit less stable. Many FOSS projects have very rapid development cycles. Downlevel code is not necessarily more stable. E.g., mplayer, WINE.

Mepis is bloated and going to an Ubuntu base makes it worse. No one is ever going to say that Ubuntu/Kubuntu/Xubuntu aren't hugely bloated. Slackware based distros have package issues and some even still use LILO! Fluxbuntu still looks very problematic.

A very tiny Fluxbox/2.6.20 system with easy expandability still looks like an opportunity. Currently, there isn't a great "small" distro that is bigger than DSL. I guarantee that the majority of Mint/Ubuntu users aren't going be to doing Frugal installs and learning the "way of DSL." Would some of them be willing to use a tiny DSL-N in traditional Debian mode with Synaptic against the Ubuntu repositories in order to utilize that "obsolete hardware?"

FWIW: Following up on what I wrote to dougz about Debian-based distros using pkgsrc... I forgot about bluewall. Lite version ISOs are under 100 MB (latest=93 MB).
http://bluewall.es.gnu.org/

That would be a very minimal "let's just get the base up and running and then build the packages" install like BSD, not a live CD with various installation options like DSL-N. So it's not exactly reinventing what's already being done.

I'll try to play around with DSL-N/Knoppix this weekend between March Madness and see what I can do. Anyone have any interest in this?

Quote (dougz @ Mar. 13 2007,19:44)

I don't want to take away the things that make DSL/DSL-N what they are. However, I think that there would be many more users if it was radically easier to add desired current packages, even if these packages don't particularly appeal to the majority of current DSL-N users. (If they want to "waste" their memory on "bloat" -- more power to them! Personally, I'm happy two "waste" the memory for K3B.)

As long as that can be accomplished without bloating the initial ISO. That's the catch-22. All of a sudden, DSL-N is Fluxbuntu with an extra installation option.

Quote

I'd submit that Debian Stable is a bit too stable for most folks. YMMV.

I agree that most people are lured into the "need" for the latest, greatest, most current release. Nevermind they're only doing the Linux equivalent of trying to make Win98 boxes run Vista. I also agree Debian's packages are slower to make it to stable than what other binary-based distros do. I'd rather err on the side of caution than take my chances with an unstable (general sense of the word) system.

Quote

Mepis is bloated and going to an Ubuntu base makes it worse.

That's why I'm sticking with MepisLite 3.3.2 and Sarge repositories.

Quote

Slackware based distros have package issues and some even still use LILO!

Because Slack still uses LILO by default. Blame Pat!

Quote

Fluxbuntu still looks very problematic.

It's a work in progress, much like DSL-N. Marathons don't start at 26.199 miles, they start with that first step at 0.00.

Quote

A very tiny Fluxbox/2.6.20 system with easy expandability still looks like an opportunity. Currently, there isn't a great "small" distro that is bigger than DSL. I guarantee that the majority of Mint/Ubuntu users aren't going be to doing Frugal installs and learning the "way of DSL." Would some of them be willing to use a tiny DSL-N in traditional Debian mode with Synaptic against the Ubuntu repositories in order to utilize that "obsolete hardware?"

First, "obsolete" is relative and so is the need for a 2.6 kernel -- especially as it relates to what is and isn't obsolete.

Second, wanna know what you're asking for? It's Fluxbuntu. It has the base required for using unstable and testing repositories, it's stripped down considerably from Ubuntu's desire to load every available byte onto a 700 MB CD, it installs like Debian/Ubuntu, can be used as a live CD, and so on. That's the direction you're asking Robert and John to take DSL-N. And I think it's already being done and it's already about the same level of maturity.

I think lucky13 and I have pretty well done this topic to death. No sense spending much more time on it.

Quote

... wanna know what you're asking for? It's Fluxbuntu. It has the base required for using unstable and testing repositories, it's stripped down considerably from Ubuntu's desire to load every available byte onto a 700 MB CD, it installs like Debian/Ubuntu, can be used as a live CD, and so on. That's the direction you're asking Robert and John to take DSL-N. And I think it's already being done and it's already about the same level of maturity.

I've run both DSL-N and Fluxbuntu. Yes, Fluxbuntu is stripped down, compared to Ubuntu, but it still contains some idiosyncratic bloat of its own. Very much driven by the vision of a very few people. I'd disagree that Fluxbuntu is anywhere near the maturity of DSL-N. It's also far bigger & slower. Current iso is 309 MB. Fluxbuntu DL page.

Fluxbuntu forums have very little activity. Last code 27-Sep-2006. No Feisty alpha/beta yet. No roadmap. No timeline.

I'm beginning to wonder if it isn't going to just grind to a halt, like Ubuntu Lite. Why in the world was Fluxbuntu to be a full, multi- platform distro instead of a fluxbuntu-desktop Ubuntu package (added to the lightweight, no-GUI Ubuntu Server install) that included scripts to shutdown daemons and uninstall bloat?

PPC, AMD-64, IA-64 support? Alternate installs, too? Currently seeking artists, but no new code since 27-Sep? Overreaching, with precious little progress to-date...

Fluxbuntu is only superficially like DSL-N. DSL-N is built on Knoppix, as well as Debian, technology. DSL-N allows Frugal, while Fluxbuntu doesn't. Also UNC, UCI, etc. DSL-N was looking like the very best bet for a truly mini 2.6. distro.

Quote

I agree that most people are lured into the "need" for the latest, greatest, most current release.

Gotta disagree with you. If the users want something, whether you or I personally agree, their desires should be taken into consideration. If a user wants a highly efficient DSL-N but then wants to install the latest & greatest mplayer, fine. They may need a feature you or I don't. It's all about choice! More DSL-N users makes for a more vibrant DSL-N user community. The user community is one of DSL's strengths and the huge Ubuntu user & developer community is also a huge advantage to all users of Debian-derived distros.

Quote (dougz @ Mar. 14 2007,17:13)

Far bigger? Compare apples to apples. It has the latest versions of various apps, which I recall you said DSL-N should provide. That accounts for a lot of it. To do what you've recommended, DSL-N will at least double. And if you mean that DSL-N should only provide more recent apps in the repository, you can start adding some of the extensions you want -- that's a community effort, not strictly the realm of the developers.

Quote

Fluxbuntu is only superficially like DSL-N. DSL-N is built on Knoppix, as well as Debian, technology. DSL-N allows Frugal, while Fluxbuntu doesn't. Also UNC, UCI, etc. DSL-N was looking like the very best bet for a truly mini 2.6. distro.

You want to use Ubuntu repositories in lieu of Debian. It would probably be easier to rebuild DSL-N from scratch on Ubuntu instead of Knoppix so that would be far more seamless.

If DSL-N remains Knoppix-based and more real Debian- than Ubuntu-oriented, let users decide how they want their systems to be -- just like now. You're not limited to using old stable or stable pools. Testing and unstable can mixed and matched, e.g.:
http://jaqque.sbih.org/kplug/apt-pinning.html

My computing needs are a little too critical to do that, but others who want to be on the bleeding edge can still run testing or unstable, or build the latest from code.

UNC/UCI extensions aren't suitable for the target audience -- traditional hard drive install -- to whom you want DSL-N to cater. UCIs can work on a hard drive install, but in a way that's totally counter your "lazy" and easier ethic.

Quote

I agree that most people are lured into the "need" for the latest, greatest, most current release.

Our disagreement isn't over whether users should be able to add whatever applications they want. We're both on the same page: anyone should be able to configure his or her system however he or she sees fit. Our disagreement is over the execution of that and whether it's the job of the user or the developers. I think the developers should be focused on providing a stable base rather than the minutiae bells and whistles users add anyway.

One can already configure or re-configure and extend (and even remaster) DSL/DSL-N or any other distro however one sees fit. That's a user issue, and it's only a developer issue to the extent that developers can make things easier or more difficult. I think the underlying philosophy of DSL/DSL-N makes it easy enough. There's also the community support here in the forums and wikis. It's a well-documented project.

Re the small group of Fluxbuntu developers: How many developers are running the show with DSL/DSL-N? I don't judge things on that basis. It only tells me a larger development group may be faster or slower than with one or two developers -- take a look at Gentoo's growth when it was headed by one person and look at what having too many developers has done to it since he left.

Finally, you mentioned Mint's addition of certain things to "ease" user experience. Some of those things are closed-source. Dittos for Ubuntu's use of proprietary drivers in the past (though they recently announced the next release will be totally open source). Do you think DSL-N should include any closed-source code -- drivers, Opera, etc. -- in the name of keeping things easy and functional, or should it stay true to the foundations of open source in general and Debian in particular?
http://www.debian.org/social_contract

Next Page...
original here.