Moving Forward - What's Your Desire?
Forum: User Feedback
Topic: Moving Forward - What's Your Desire?
started by: roberts
Posted by roberts on Mar. 30 2007,18:13
Just to gauge the DSL user community, please vote. Results are not binding. This is for my own interestGiven that the DSL community has many members who can help by compiling modules, is it time to do some serious updates?
Given limiited development time and developers, only one direction, only choice can be undertaken.
What direction would you like to see for Damn Small Linux
Be aware the ramifications of your decision:
2.4.34 is still being actively maintained. Many new features are being backported to 2.4 series.
2.6.20 is much larger both in size (no syslinux version) and runtime (many smaller ram machines may be left out)
A tiny MyDSL core may not be useful upon first boot and/or without high banwidth net connection. But with perhaps good introductory documentation it may be the choice.
On the otherhand, more or newer apps will be larger, and merging the two, DSL & DSL-n will also be much larger, thereby leaving out smaller ram machines.
Please give some careful consideration before you vote and feel free to discuss and leave comments below.
Thanks,
Robert
Posted by lucky13 on Mar. 30 2007,18:51
I gave this subject a lot of thought when I was considering doing a HD install of DSL-N plus pkgsrc -- how best to appeal to those who want bleeding edge apps while not leaving people who want stable-oriented systems out in the cold.I voted for "Remove many apps providing a tiny core based on using MyDSL 2.4," but my preference is for doing that by updating the 2.4 kernel to 2.4.34 and updating the base libraries and a small group of applications. I'd hate to see the legacy support of the 2.4 kernel DSL has retained lost.
The included applications should be just adequate for leaving the CD useful for running live while increasing the focus to extending DSL installations (frugal, USB, and HD) via MyDSL. Updating the base libs (GTK2, for example) will allow users to run more recent applications. For development, I think upgrading applications will be more efficiently done over longer periods of time via extensions than doing it via the CD. For users, it would mean an easier process of making and using extensions of recent versions of popular applications.
In a nutshell: freshen it up a bit and make it even more modular.
Posted by mikshaw on Mar. 30 2007,19:50
I haven't yet voted.I agree with lucky13 mostly, wanting to see DSL become more modular as far as applications go. This may make it possible to include more hardware support and other kernel modules in the base which may not be so easy to use as extensions (assuming it actually is more complicated to add hardware support via mydsl?). It could be possible to support a greater variety of fake modems, video cards, input devices, and other hardware while maintaining the 50mb limit.
Perhaps it could even support a more flexible X system out of the box, for those users whose monitors look like crap at 60hz.
At this time, I'm undecided between 2.4 and 2.6 kernel. I suppose if newer hardware support (sata, for example) is backported to a useable and stable 2.4 kernel, there probably isn't a great need for 2.6, but as I still have never used a 2.6 kernel in any distro, i couldn't say from experience what other benefits can be had with it. I have so far been satisfied with 2.4 with the hardware I have.
Posted by MakodFilu on Mar. 30 2007,20:11
Upon discovering DSL, at first I (as many) wanted "this and that" application included (not so many). To be fair, DSL matched my tastes to a great degree: my Debian box had almost the same applications (should be said "lack" of bulky applications).Then I learned to love myDSL and just keep my 2-3 extra apps around in the USB flash drive. I think the DSL core applications can solve almost any basic need as a desktop user. Most of them. If someone asks to remove some application, even if I don't use it, I wouldn't know which one.
I heard of a library implementation of xpdf, I also liked the way the Links issue was solved (if you want to browse the web, you should have web access, so why not downloading the .uci?). In fact, I think that applies for Firefox as well. Why not removing Firefox and download the .uci automatically upon clicking the xtdesktop icon?
As for the kernel, if 2.4 is still being maintained, I don't see a real need for 2.6
Posted by humpty on Mar. 30 2007,21:09
Would like to see DSL keep it's identity (it is not puppy).But what to choose? trim down the apps? Yet it should still
keep up with the times? oh i hate these decisions.
I would eventually opt to cut back on the apps and go for kernel 2.6. I've heard good things about 2.6;
< http://kniggit.net/wwol26.html >
Posted by lucky13 on Mar. 30 2007,22:28
I've written a more detailed explanation of the views I expressed earlier on this subject, with more about how I think the changes I advocate would benefit users and developers.< http://lucky13.blogsavy.com/dsls-future/ >
Posted by torp on Mar. 30 2007,23:39
don't pull DSL out of it's niche.torp
Posted by lucky13 on Mar. 31 2007,00:08
humpty:| Quote |
| I would eventually opt to cut back on the apps and go for kernel 2.6. I've heard good things about 2.6 |
I've read this before. It was written about four years ago (per the 2003 date on bottom of the page), before a lot of the features he notes are native to 2.6 were backported to 2.4: SATA, NPTL (which he noted had been done by Red Hat in RH9), etc. ALSA modules are also available for 2.4 users.
I know there are things many users immediately would appreciate in 2.6 over 2.4, like handling of hot-plug events via udev (a need or a convenience?) and (maybe?) better X response. The trade-offs, though, for DSL users in particular are loss of use of older devices for which support is dropped from 2.6, significantly larger kernel size (all these internal improvements come with a steeper price), and I'm still not sold on speed issues (X, etc.) as it relates to older hardware -- which has been one of DSL's real strengths -- because there's an effect of balancing things out when you compare resource demands required by 2.6-only daemons like udev/hald against the improvements in more limited scenarios. I know that won't apply to people with fairly recent hardware and computers 256+ MB RAM for which 2.6 was intended. How significantly will affect users with older hardware? See some of the comments in this thread for examples of what I mean:
< http://lwn.net/Articles/140975/ >
Since 2.4 is still in active development and features continue to be added/backported while not deprecating legacy support, I don't see a need for DSL to switch. I'm interested to see the responses of other users and their (specific) reasons for wanting to switch to a 2.6 kernel.
Posted by mikshaw on Mar. 31 2007,01:39
As far as I see it, support for very old hardware should remain as long as there is a significant amount of that hardware still in use. This was, after all, one of the reasons DSL was born.If much of that support is lost in a 2.6 kernel, I don't see any reason to go beyond a newer 2.4 kernel.
I'm sure most people use systems that are newer than 3 or 4 years, but most is not all. As has been said by others, it's a waste to leave the older machines behind if they can still be useful, and keeping DSL updated and improving while retaining support for these machines is something that practically no other distro can claim.
Posted by Juanito on Mar. 31 2007,05:27
I cannot see a difference in speed between DSL and DSL-N on my laptop, but I can see a big difference in power saving features (cpufreq, laptop-mode-tools, etc) and bluetooth/irda that make the laptop much more useable in DSL-N.On the other hand, I did not even think about using DSL-N on my ancient desktop (1997), which works perfectly under DSL and doesn't need power saving or bluetooth/irda/etc.
I believe the focus should be more on getting applications to work better - it would be great to be able to use a bluetooth headset with Skype for example - and to make the base system more easily upgradeable.
At the moment, dpkg/apt-get doesn't recognise that Perl is present in DSL, the libc6 library needs upgrading to install almost anything, etc, etc - all this makes installing a new application that much more difficult.
Anyway, whichever way we decide to go, I'll be happy to continue using DSL and DSL-N - thanks for a great, and different, take on Linux.
Posted by curaga on Mar. 31 2007,13:19
I also voted for upgrading DSL to 2.4.34.Slightly off-topic, but Roberts, how is DSL actually done?
I mean, is it a dedicated HD-install somewhere, which you modify, and when happy, just do a mkisofs in the root dir?
Posted by roberts on Mar. 31 2007,14:30
| Quote |
| Slightly off-topic, but Roberts, how is DSL actually done? I mean, is it a dedicated HD-install somewhere, which you modify, and when happy, just do a mkisofs in the root dir? |
No. Over the years I have modified several scripts that allow much automation of the currently four DSL images based on, what I call, the working directory method of remastering. This allows me to perform such from any published DSL CDROM and a mounted ext2 partition both direct access and chroot.
I usually test new additions on a live or frugal environment.
The only time when I rely on a full traditional hard drive installation is when I upgrade the kernel and modules.
This I always do at runlevel 2.
Both of which, remastering and kernel building, can be on the same machine or networked machines.
I never want to have "all my marbles" tied to a particular machine.
Posted by curaga on Mar. 31 2007,14:36
Thanks. That clarified it.
Posted by clivesay on Mar. 31 2007,20:05
I raise my hand for 2.4.34 on a tiny core. There is a fad in Linux right now to take large distros and make a 'Lite' version. These are already covering the 2.6 kernel niches. I don't see a reason for DSL to add to that list. If RS went this route then I guess the question would be how small to make the core and what version of Debian to base it on? I'm talking over my head here.
I have always felt that a small core with apps added over unionfs and uci is the strength of Robert's DSL development. Why include an old Firefox in the base when someone is going to immediately grab a Firefox uci to get a newer version? That's just an example.
Make the core small and turn the apps over to the community. If someone wants to run kdelibs over unionfs, then let them do it.
Chris
Posted by newOldUser on April 01 2007,13:48
Hey lucky13, I like your blog post on DSL's future. You make some good points.I run DSL two ways. One, as a primary operating system (frugal install) on older hardware that I want to use as a Server. I add a couple of MyDSL packages and poof, I have my music/web/webcam/samba/file/you-name-it server up and running. I find it a great way to re-use old hardware. I like that it has a low resource GUI interface so I don't have to remember all the command line mumbo-jumbo when I'm making changes.
Secondly, I run DSL as a temporary operating system (via usb stick or live CD) on more current hardware. Usually, I only do this to demonstrate Linux to others or do some repair work when the primary operating system on that machine will not boot.
I think if people are looking for a primary operating system on newer hardware they'll probably go with something like Ubuntu, SUSE, Fedora, etc
Why did I first pick DSL? I was interested in learning Linux, I had some old hardware to experiment with and I was tired of downloading large distributions only to find out that they didn't work with my hardware or they were too complicated for me to understand. DSL was a quick download and the Live-CD feature let me get Linux up and running in no time. After that the forum community support and the MyDSL modules kept me around.
My vote:
1 - Stay with the 2.4 Kernel, with updates to .34, for at least another year, then re-evaluate the user base... eventually “newer” hardware becomes “older” hardware.
2 - Remove applications as needed to make space for more hardware support. Firefox was a real slick thing to have in a 50mb distribution but if I can download it as an add-on and still have Dillo in the base, that's fine.
3 - Don't worry about having to fit on a mini-cd. Heck, I think 64 meg usb sticks are easier to find and cheaper then mini-cd's in some locations. Don't let the size get too large but I don't think many people will leave DSL if it becomes a 60 or 70 meg download.
4 - Don't kill yourself putting out release updates. If you can do two or three releases a year, that would be great. I would hate to see Robert/John get burned out. I know it's a labor of love for them but I don't want them to be under the pressure of a new release every other month, unless that's what they want.
Thanks for the great ride.
Posted by lucky13 on April 01 2007,15:40
| Quote |
| Hey lucky13, I like your blog post on DSL's future. You make some good points. |
Thanks. A couple things I didn't elaborate too much on were points you've raised:
| Quote |
| 2 - Remove applications as needed to make space for more hardware support. |
I didn't write much on the issue of hardware support aside from commenting on what's been backported. There will be more hardware supported in 2.4.34, so depending how the kernel is compiled (and with respect to how many modules are compiled), DSL will still work with older hardware and it should be much less of a hassle to get newer hardware working. And this should mitigate what mikshaw alluded to about modules as MyDSL extensions. Fortunately, I haven't had to use any module extensions on my computers. (Old school rocks.)
| Quote |
| 3 - Don't worry about having to fit on a mini-cd... Don't let the size get too large but I don't think many people will leave DSL if it becomes a 60 or 70 meg download.... Don't kill yourself putting out release updates. |
This goes to my focus on increased (total!) modularity. If the ISO is mostly confined to the very basic stuff -- kernel and utilities, with just enough applications to make it useful as a recovery/live CD -- there can be more hardware support in the same space or not too much more. Limiting the scope of the ISO (from currently having a little bit of everything to eventually being oriented more for installation while retaining sufficient tools to be used for recovery or other "emergencies") also would reduce the need for frequent release schedules. The (a) kernel and utilities have a longer "shelf life" than (b) applications, so keep the two (a and b) as separate as possible.
I'm not as concerned about the size of the ISO as it relates to CDs or USB sticks as I am about how most people seem to use DSL. Increasing modularity will allow every user to select only the apps he or she wants. Someone installing to a 256 MB stick, for example, won't have a couple of those MB taken up (in the ISO) with apps he or she is going to replace via MyDSL anyway. That issue is much less a concern with installation (frugal or traditional) on a hard drive, but I think the whole separation of a smaller "base" from the apps is already consistent with the direction DSL has been going and will improve things.
I don't think anyone would go as far as I would with respect to stripping out apps (I'd leave one wm, mc, elinks and dillo, elmo or sylpheed, naim/nirc, cdw, etc., no multimedia apps, nothing fancy but still adequate for someone to restore a system quickly or to have minimal tools to get online and find help). I'll concede, though, that such austerity would reduce DSL's appeal as a live CD. I don't know how many people will get huffy if the games and Firefox 1.0.6 are removed. Those who want more on a live CD could still go through a (p)remastering process to add in whatever apps they want from MyDSL. Depending on perspective, modularity would benefit them as much as everyone else because they'd get a CD with exactly what they want.
Posted by roberts on April 01 2007,17:14
Let me state that I have no interest in making DSL a rescue tool. There are already many such distributions out there. Some, in fact, were originally based on DSL. Nor will DSL become a Debian hard drive installer. Debian already supplies that.
DSL will not compromise its Linux/Unix roots and run everything as user root just to try to accomodate those less familiar or unwilling to learn the *nix way of running an OS, or because it is easier.
At minimum DSL will retain its niche in the sea of distributions. That being not only a physically small nomadic, yet easily extendable, distribution, but one that runs well and supports very small hardware. That is the direction that I have provided to DSL with my development efforts.
Debian compatibility and traditional hard drive installations have been secondary. Again, there are many other distributions that have that as their focus.
As the developer, "I eat my own dog food". In other words, I use DSL as my primary OS. Over the years that I have maintained this OS, I have tried to keep DSL true to John's 50MB Desktop OS concept. I am also keenly aware of supporting old(er) hardware, and of being true to a nomadic well behaved guest OS, either booting natively or via emulation. I am very much drawn to the "Small is Beautiful" concept.
DSL will not become so stripped to be not useful to me upon first boot. Yet there are some applications that over time have become less useful and I too, often mydsl overlay them. Example firefox 1.0.6 and xpdf. These two could be possibly removed in favor of their mydsl counterparts, or perhaps the more capable, but larger versions, be added but then exceeding the 50MB limit. Most of the other applications present in DSL do not have such issues and I find that they serve their function very well. If you wish to comment or suggest others feel free to do so.
Recently, During this v3.3 RC cycle, I tried to remove the non-boot scsi modules and offer them as download extensions. Really, how many of us still have or use old sccsi drives. In fact, I do. Daily, I use an old scsi tape drive to backup my development system. Yet, I was still thinking to remove them. Doing so is a savings of almost 2MB. Yet this seemed to be rejected.
I also tried to remove the netfilter modules and package them with the iptables.unc. Again, we are carrying modules that cannot be used without the appropriate program, iptables, a mydsl extension. Yet there were complaints.
There is an old saying, it is easy to give, but so hard to take away; even when you still make it available.
As for my recent push to publish many many enhancements for DSL over the last several months, well that will become very apparent later this month when a major annoucement will be made.
I appreciate the comments and voting. After v3.3 is final, looks like I will be looking to implement 2.4.34. Likely our usebase is here because of DSL's support on smaller and/or older hardware, thin clients, or embedded devices. I will continue to strive to keep up.
I look forward to continued interaction with those who share my interests and enthusiasm for the little OS that could ( and does!)
Robert
Posted by Juanito on April 01 2007,17:26
| Quote (roberts @ April 01 2007,17:14) |
| After v3.3 is final, looks like I will be looking to implement 2.4.34. |
If that's the decision I'll be fine with it - 'would be kind of good to implement DSL-N 1.0 first though...
Posted by lucky13 on April 01 2007,18:18
| Quote |
| Let me state that I have no interest in making DSL a rescue tool. |
I don't know if this was meant generally to all or specifically to something I wrote. I didn't mean to suggest that DSL should be a recovery tool, but that it should have at least enough utility as a live CD to do common and simple tasks like browsing, e-mailing, getting someone's nuts out of the fire with respect to backing up files, etc., so that it wouldn't be stripped so austerely that it would effectively only be an installer. I didn't mean it should be packed with a full alphabet's worth of filesystem tools beyond what DSL already has for ext2/3. You're right, I can do that already with my remaster or other distros that do it by design.
| Quote |
| After v3.3 is final, looks like I will be looking to implement 2.4.34. Likely our usebase is here because of DSL's support on smaller and/or older hardware, thin clients, or embedded devices. |
Updating the kernel should expand DSL's usefulness on more newer hardware without disenfranchising those of us with older hardware. I know you have a lot of things to balance in deciding what stays or goes in keeping things consistent with DSL's basic concept. I think you're doing a great job at consistently improving it given the 50 MB ceiling.
Posted by roberts on April 01 2007,19:44
I agree that balance and being on the conservative side has kept me focused to keep improving what we have. I think the community and John keep me in check. I can be just as radical as the next guy. I do tend towards what interests me. Compiling kernels and modules are a boring task.For example, my recent posts in water cooler on self contained applications led me to proto-type a ROX desktop running in DSL
I stripped away xtdesk, emelfm, fluxbox, and jwm. Slightly modified UCI to become well behaved ROX Application Directories, then with an old Oroborus WM and Rox panel I had a menu-less ROX drag-N-drop desktop with MyDSL UCI's in the Apps folder running in DSL.
Now that's radical. Yet it is interesting. What is old RISCOS, is new again. But running in a small Linux environment. Much smaller than the ROX Desktop/Python currently being developed. It is newbie friendly and no, none, nada, package management needed! But could you adapt to a data driven desktop and away from menus galore and application (icon) driven desktop that is the norm for Linux and Windows systems?
Posted by lucky13 on April 01 2007,20:50
| Quote |
| I stripped away xtdesk, emelfm, fluxbox, and jwm. Slightly modified UCI to become well behaved ROX Application Directories, then with an old Oroborus WM and Rox panel I had a menu-less ROX drag-N-drop desktop with MyDSL UCI's in the Apps folder running in DSL. |
That's not radical, that's (bleeping) cool. Did you set a MIME type for UCI and then a run action to load and mount them directly or set run action to a script to mydsl-load and then mount/unmount them? I'm curious how you modified the UCIs to become "well behaved."
Posted by humpty on April 01 2007,21:46
| Quote (roberts @ April 01 2007,21:14) |
| .., well that will become very apparent later this month when a major annoucement will be made.. . |
(dramatic music)
:-o
Posted by lucky13 on April 01 2007,22:40
Robert: Now I think I know how you set up your UCIs as AppDirs (not sure how you modified them, though). Don't tell me! I'm going to try after the game tonight or tomorrow and see if I can figure it out.
Posted by WDef on April 01 2007,23:09
Hi all,1. John's original concept and Robert's fine exposition of it is a proven, winning formula viz.
Functional, fun desktop + 50MB cap = damnsmalllinux
Everyone loves it. So, whatever else may happen, I'd hate to see the "dsl classic" side of the project simply dissappear - that would be foolish. And people need a single cd dsl like it is ie boot it and use it, no fooling around with loading modules. Whichever kernel and whatever other directions are pursued, I really think that's what defines dsl.
2. HOWEVER, that doesn't mean there isn't room for experimentation on the side. I absolutely *love* the sound of the Rox Desktop dsl. Even if it goes nowhere Robert, please post your efforts somewhere so we can play with it.
Changing kernels: I personally would like to see a more advanced 2.4.34 50MB dsl, but I'm mindful this resulted in alienating the old hardware users last time (2.4.31). How many of them are there? Are they a small percentage of users? Can this be worked around?
Finally: I thought the dsl classic + dsl-n model was a very good one, providing two development paths for each kernel and moving ahead from dsl's old glibc and gtk1. These are more of a barrier to new apps than the actual kernel. I know there were resource problems, and these won't go away for another revision of the dsl project (I may take this up with Robert privately as I have some ideas). I do think however dsl-n was alpha released a little early. To garner more interest, it needed things like a few essential extensions ready to use, and a clear demarcation between dsl classic extensions and dsl-n extensions. There still is a lot of interest and it would be really great to see dsl-n revitalized.
I guess obviously the limit is really how much in the way of resources Robert and John and the community can muster.
The modular approach sounds great and I would like it myself very much, but *not* if it's at the expense of a tight full desktop 50MB distro. That needs to remain no matter what. The people wanting the modular version are probably more advanced users and hobbyists, but dsl picks up a large number of totally new linux users all the time. They need to be able to boot a single cd and see a working distro.
I could see a modularized dsl being an interesting child project perhaps for others to pursue if it's outside the resources of John & Robert to implement.
If economies of scale can be created so the work that goes into one arm of the project easily transfers to another, then great.
Anyway enough out of me tonight.
Posted by Jason W on April 02 2007,03:36
I would like to one day see the libraries/apps updated. That would make compiling apps easier. But that don't have to be anytime soon. And Debian compatibility is not that important to me, it at best would be a nicety. There are all kinds of alternatives if we want the latest and greatest. Debian-Live will give us a Sarge or Etch live cd. And there is always Knoppix. But only DSL works with old hardware (64 or less mb RAM). Being a Woody fan, I appreciate the fact that DSL has remained with the Debian Woody base since (I think) the beginning. And in a world of constant clamoring for the latest packages, DSL is rare haven for those of us who like to stick with what works. But we also get treated to some of the latest technologies in live cd's. So in other words, stay the course, and keep up the good work Robert.
Posted by humpty on April 02 2007,08:25
there's no reason why the apps can't be seperated and then distributed along with the iso is there? may put them in one file 'standardapps.uci/unc?' on the root dir of the cd?
Posted by clivesay on April 02 2007,13:32
| Quote (humpty @ April 02 2007,03:25) |
| there's no reason why the apps can't be seperated and then distributed along with the iso is there? may put them in one file 'standardapps.uci/unc?' on the root dir of the cd? |
Humpty is spot on to my point. Make every app possible a uci or unc. The DSL devs can decide what apps they want to package with the live cd. To the new user, it's completely transparent. They boot a live cd and the apps are there and working. To the advancaed user, it's a modular distro where they can add/remove the uci/unc that they want and then do a Make My DSL type livecd remaster.
To me, THAT would be a trick distro. It can truly be as small and large as you want and all you have to do is add/remove unc/uci files to suite your needs.
If the core could be made small enough, you could possibly run the core toram in 64 or 96mb ram and still be able to add additional apps via uci/unc. Imagine reviving old hardware if you could run DSL toram on a P1 with 96mb ram! Wouldn't that impress your friends.
I hope my previous post didn't suggest debian compatibility. I just meant the files in the core. Now you have a hard time making some apps into extensions because of the base libs. If the core can be made so it's compatible to whatever version of debian you wanted to pull your apps from, that would be cool. Again, this part is way over my head.
Chris
Posted by roberts on April 02 2007,16:02
Actually I built DSL-N in that very manner. I maintain a tiny core and all the applications are added to the ISO during the mkisofs step. Of course the resulting ISO never revealed this fact. I very much like this configuration. It also allowed for separation of development efforts. Where I usually concentrate on the tiny core.
The only concern that was expressed to me, was that releasing this configuration would likely result in many competing clones based on my tiny core. Many of us recall that we once had such a shadow.
I have not attempted to break apart DSL is such a manner. But that could be the decision and priority in moving forward.
It is true about the lower memory usage. In fact, I have also proto-typed a PXE version of DSL. This version has KNOPPIX image embedded in the initial ramdisk. This makes booting much easier, as no other boot time drivers need to be loaded to try to find the KNOPPIX image. That means no more "can't find knoppix image - dumping to a minimal shell" message. Therefore, the lower initial ram usage, the better for this particular configuration.
So many choices, so many roads still to travel...
My decision is still not clear.
Keep voting and posting your comments and suggestions. I am following your posts closely.
Robert
Posted by roberts on April 02 2007,16:12
| Quote |
| 2. HOWEVER, that doesn't mean there isn't room for experimentation on the side. I absolutely *love* the sound of the Rox Desktop dsl. Even if it goes nowhere Robert, please post your efforts somewhere so we can play with it. |
Send me an email and I will answer with a link. I wonder how much interest there is for this. This environment is actually much closer to my overall philosphy and can fit well with what I have developed. A DSL Rox Lua Desktop sounds like a fun environment both to use and to develope.
Posted by lucky13 on April 02 2007,16:58
Robert:| Quote |
| I wonder how much interest there is for this. |
Please count me in. I haven't had time to fiddle around with setting up UCI AppDirs this morning -- just going through a few ideas about how to do it while stuck in traffic. Will you tell me what you did to get them to "behave"?
clivesay/humpty: That's what I was alluding to about modularity and (p)remastering.
Posted by clivesay on April 02 2007,17:07
| Quote (roberts @ April 02 2007,11:02) |
| The only concern that was expressed to me, was that releasing this configuration would likely result in many competing clones based on my tiny core. Many of us recall that we once had such a shadow. |
In the Linux world you're always going to have this type of thing, as you know.
If you have a truly tiny core, what is there to really clone? All apps are 'modules' (unc/uci) so the core doesn't have a lot of use unless these apps are available to be attached to it. I'm just talking off the top of my head here.
I'm thinking the core is only what it takes to boot Knoppix and all of your Lua magic. Some of that may even be able to be modularized. With this approach you have tight control over the core and maybe some base extensions then the community carries the weight of developing all the other app extensions, just like what happened when Mydsl first emerged.
In this scenario, you could still have a "DSL Classic", if people wanted, where any innovations like new lua scripts could be backported to "Classic" whenever possible.
Actually, as you make updates you could backport them to "DSL Classic" by packaging them into uci/unc. That would keep people from having to d/l the whole 50mb cd again for most updates.
Just brainstorming here. RS you are an innovator so I will trust your direction.
Chris
Posted by mikshaw on April 02 2007,19:10
A thought that just now re-occured to me, and something i was playing with several months ago, is the possibility of modifying myDSL in such a way that it is not tied to the window manager, that is it doesn't create a fluxbox menu item which is then converted to a jwm menu item. I haven't worked out a solution that would make a menu item still useable, but I think if it could be done it would make things more convenient both for a tiny DSL core and for those users who want to add or remove window managers. Currently the user still needs .fluxbox/mydsl.menu (i think) even if he's remastered DSL without Fluxbox, and having a generic (non-wm-specific) menu *might* make it easier to include the mydsl menu into an arbitrary window manager.The main problem with this, apart from the time and effort to modify mydsl, is the existence of a fluxbox-specific menu item in most of the existing extensions. A workable option, though, might be to keep the current fluxbox syntax in the extensions and simply check for the existence of a mydsl.menu file before writing the menu item. Maybe this check is already being done, though.
In any case, it's something to think about if this tiny DSL core is actually put into play at some point. Making DSL easier to strip down will likely result in more people stripping it down =o)
Posted by clivesay on April 02 2007,19:26
Mik, So something like a menu.lua that makes the menu OS independent? That's an interesting option. Then you really could get down to a core where even the WM's are optional but you can still build menus.
With flexibility comes complexity.
Chris
Posted by lucky13 on April 02 2007,22:24
Mikshaw:| Quote |
| I haven't worked out a solution that would make a menu item still useable, but I think if it could be done it would make things more convenient both for a tiny DSL core and for those users who want to add or remove window managers. Currently the user still needs .fluxbox/mydsl.menu... |
Free desktop standards (XDG) require a centralized XML file from which every compliant window manager can adopt (via "include") the entries in that file. It's just a matter of centralizing menu entries for MyDSL extensions to one file that can be included across the board by adding one entry to the menu of whatever window manager a user chooses. During my search I found a link about how Puppy has already transitioned to do this (second link). It would affect DSL (and be implemented) similarly.
< http://standards.freedesktop.org/menu-spec/menu-spec-latest.html >
< http://puppyos.net/blog....-002857 >
Posted by lucky13 on April 02 2007,22:26
clivesay:| Quote |
| With flexibility comes complexity. |
With standardization, though, comes uniformity, ease, and utility.
Posted by roberts on April 02 2007,23:08
I am aware of the centralized menu system. However, the older gtk1 versions of most WM do not support this. It wasn't around then. Specifically the fluxbox currently in use. New WM's generally mean Gtk2, or at a minimum more bloat. It is always easy to give in and bloat up. It seems to always come down to writing some custom code or using/assembling larger pre-made subsystems. While it is good that standards have emerged, there is little effort to keep size to a minimum let alone to still support gtk1.I can tell you the current fluxbox is most favored for its size and functionality.
Most of my challenges were to make disparate incompatible systems work well together via custom code. The combination of xtdesk, fluxbox, and emelfm, have all given me fits over the years. Yet they represent a very small working environment that have served us well.
Posted by humpty on April 03 2007,03:51
id have to agree with roberts, and there has to be some sort of identity left in the core.
Posted by curaga on April 03 2007,16:22
Then there is the LinuxFromScratch ideology - Build everything yourself.It seems that most of stuff in DSL is from debian repos, meaning no optimization. If one would compile some programs, maybe tinyx and fluxbox, with
| Quote |
| -march=i486 -Os -fomit-frame-pointer -s |
And also optimize the kernel for 486...
Posted by WDef on April 03 2007,17:53
The more I think about it, the more I like the idea of:1. A 2.4.xx dsl "classic".
Plus
2. The modular approach but applied to 2.6.xx + gtk2 ie a DIY kit for dsl-n
the latter is not contrained by a 50MB limit and could be a lego kit for rolling-your-own.
RE "shadows": imitators, clones can be a good thing. They can do testing and feed improvements and bugs back upstream. They can test the market. They can bring users to the "original and best". This is how the GNU GPL world works, isn't it? Torvalds: "Linux is evolution not intelligent design".
There's room for a lot of microdistros, there are already a few, and DSL dominates without any effort whatsoever.
Also: corporates might use such a kit for quick solutions, or at least for fooling around. They have access to a lot of resources, and could be encouraged to support the dsl project both cash and in-kind, and in expertise.
Consider the issue (on a different scale!) with CentOS and RHEL.
CentOS is probably good for RHEL - it provides a bigger base, and RH is always going to sell support to coprorates.
And Knoppix isn;tthe least bit threatened by all its spinoffs in clones, rather, they revel in it. I think they are probably quite fond of dsl.
Posted by stoneguy on April 04 2007,00:32
I'm not voting, because that's not how I would make a choice.DSL is intended to run on "older" machines. But there are degrees of age.
Let's divide home PCs into eras. The most useful stratification is "What Microsoft O/S was it delivered with?" So we can talk about
pre-Win9x machines
Win9x machines
WinXP machines
Vista machines
What processor the PC has doesn't mean much, as long as it's some flavour of Pentium. If you're slower, you'll just have to be more patient. The critical factors in these machines is the RAM supplied. Win9x machines were usually sold with 64MB or better. XP machines had at least 256MB, and then salesperson tried to convince you to upgrade to 512. Cheap buggers like me didn't
Another axis of differentiation is "newer-vs-older hardware" machines. Newer can have SATA, bootable USB2.0, maybe Bluetooth, Firewire, etc. Older will be IDE, maybe USB 1.1, DVD writer, etc.
If your system falls into the newer group, I'm not sure why you want to runDSL. You folks can run any Linux you want. Why not one with glitz and glamour?
Next, I'm going to define something called the Modern Internet. On it, participants expect to be able to access images and sound in eMail and at Web sites, view Flash animations, read documents in PDF format, and upload from their point-and-shoot digital cameras. (I'm not even sure they'd use Linux if it can't run the software that came with their cameras.) They expect the ease-of-use and universality of IE and Outlook Express, which pretty well means delivering Firefox and Thunderbird. And the participants I'm talking about are not geeks. I'm referring to Joe six-pack-of-Geritol® here, who isn't up to loading ucis or dsls or whatever dang fool thing to get to the point where Windows is out-of-the-box.
So what systems should DSL target, and what should it deliver?
My inclination would be to write off the pre-Win9x machines. Although there's always "because you can" bragging rights, those machines are too outdated for use on the Modern Internet. That leaves DSL's niche as the Win9x older hardware systems.That's where it'll really shine and serve a useful function.
And whatever it takes to deliver the Modern Internet, that's what DSL neds to deliver.
And one more thing. Lose the 50MB restriction. That's just those bragging rights again. I don't know what it's like elsewhere, but in Canada's largest city, I can't expect to walk into any computer store and walk out with credit-card CDs in my hand. I can't even buy the minidisk format, except for DVDs for cameras in places that cater to the photography crowd. And the final insult is that the small media has always been a lot more expensive than full size.
I've finally euthanized my Pentium 133MMX 64MB system, so I doubt if I'll be lurking around these parts much any more. My current low-end system is a 400MHz Pentium with 384MB RAM. That's sufficient to run the Ubuntu family.
Posted by mikshaw on April 04 2007,02:36
stoneguy: I started responding to individual parts of your post, but decided that would be more argumentative than useful (and was taking up too much time), so I'll just say this...I disagree with practically everything you just said. =o)
Posted by WDef on April 04 2007,14:18
Interesting analysis, Stoneguy. I always been dead against losing the 50MB restriction, because (as I've said) that is what *defines* and has defiined dsl. That is what creates the discipline in its design and motivates the preference for light, tiny apps. Otherwise it just becomes dsl-n or another minidistro. You might as well move to Vector or something. It's no longer much to do with creditcard cds and hasn't been for a long time, that's just a remnant in its marketing. It's to do with light and fast and small, and being able to do toram boots on 128MB ram.
This is why I like the dsl classic + dsl-n type of model: the latter loosens the restrictions and moves into gtk2 and fatter apps, but we don't lose the heart of dsl in classic dsl. A modular approach would also be much more useful in dsl-n.
| Quote |
| If your system falls into the newer group, I'm not sure why you want to runDSL. You folks can run any Linux you want. Why not one with glitz and glamour? |
I run dsl all the time on a 1.87GHz Centrino 512MB ram as a toram boot. Why? Because it absolutely screams and I love speed! And I like minimalist apps in the unix tradition. A number of other users also use dsl on newish machines for the same reason.
For 'glitz and glamor' I barely tolerate Fedora, which has it all, and other distros in the recent past. In terms of desktop responsiveness, KDE crawls along like a snail on valium.
PS: I also haven't voted because none of the options exactly match my desires.
Posted by torp on April 04 2007,19:37
| Quote |
| I've finally euthanized my Pentium 133MMX 64MB system, so I doubt if I'll be lurking around these parts much any more. My current low-end system is a 400MHz Pentium with 384MB RAM. That's sufficient to run the Ubuntu family. |
good riddance as far as i'm concerned. i like my dsl small and fast, and modular. scalability is the key.
torp
Posted by jpeters on April 07 2007,03:36
| Quote (torp @ April 04 2007,15:37) |
| good riddance as far as i'm concerned. i like my dsl small and fast, and modular. scalability is the key. |
I couldn't agree more. The more it looks like windows, the more bugs , reboots, freezes, etc., etc.; exactly why I'm no longer using windows.
(OTH..I have no idea why someone would choose DSL to run a recording studio...)
Posted by humpty on April 08 2007,07:36
i'm sure most of us don't agree with stoneguy's points, but then i just realized that some/most? of the users that stumble upon this site probably share the same view, stoneguy reflects a portion of pc users which other distros gladly target. i would like to think more users are looking the other way nowadays.dsl is still 'different' enough to keep it's place. the mydsl modularity ensures that. maybe years ago, only us 'geeks' could appreciate that, but these days the popularity of the U3 software has taught more people that 'lean is mean'.
likewise, the 50mb limit is surely an attractor. sure you can store more on today's pen/usb drives, but everybody knows the more room you have left over, the better. you don't need to be a geek to know that.
seperating the apps from the base is actually a striking concept. to say to the the guy on the street that you can load this software by just placing it in this directory, or to load it only when you want to, and there is a repository of this stuff, and even if you get in a mess, you just have to reboot. This goes down well too.
Posted by Unknown User on April 10 2007,15:08
| Quote |
| i like my dsl small and fast, and modular. scalability is the key. |
Exactly. Dsl is really fast on -98 AMD Athlon 600mhz and 256mb ram.
Posted by u2musicmike on April 10 2007,22:28
I voted to "Remove many apps providing a tiny core based on using MyDSL 2.4" but I wouldn't remove too many just enough to be able to update the kernel to 2.4.34. I use DSL on laptops and old desktops which had windows 95 on them. I feel that DSL is a big security improvement over windows 95 and I will never go back. My thoughts is that a new kernel will be even more secure and may support more devices like scanners. I like the 50 Meg size so I can download easier and do frugal installs without using a big hard drive. A hobby of mine is using DSL to restore old PC's and I give them to friends who need a PC.
Posted by buzzard on April 15 2007,22:42
I didn't vote, but I agree that if room is needed, apps like firefox could be stuffed off onto .uci. What I like about DSL is its modularity, and wouldn't want to lose that in the shuffle. As for kernel modules, things like imm.o (zip-plus), and like someone else said, winmodems. (needn't be included in base, just available for adding-on by users). The MyDSL files kind of remind me of mac os 8 where the extensions, left in a certain folder, were automatically added at boot. I notice there are 3 ways to add mydsl: .dsl, .uci, and .unc. Does this mean that the ultimate best way to add modules is still being searched for? I wouldn't mind it if the add-on apps were loaded via bootlocal.sh, with bootlocal.sh being loaded from a location specified in the kernel command line
Posted by damnit on April 16 2007,01:17
i like the way damn small has been designed, i started using it on version 3.01. it is ready to use out of the box and can very easily be modified for individual tastes. i voted to simply update to 2.4.34 rather than bloat damn small and upgrade apps that may not need upgrading. i
Posted by andrewb on April 16 2007,04:05
With the advent of unc extensions & auto-loading of GTK at boot-time there isn't any need to have Firefox in the base system. Dillo is sufficient for most browsing & Opera or Firefox can readily be added via unc. This would also remove the gripes from some about the age of the Firefox version in the base system as the latest version could be readily provided via an extension. Removing this bloated app would give plenty of room for a newer kernel & assicoated modules to keep the base system compatible with legacy systems.
Posted by jpeters on April 16 2007,04:38
| Quote (buzzard @ April 15 2007,18:42) |
| I notice there are 3 ways to add mydsl: .dsl, .uci, and .unc. Does this mean that the ultimate best way to add modules is still being searched for? I wouldn't mind it if the add-on apps were loaded via bootlocal.sh, with bootlocal.sh being loaded from a location specified in the kernel command line |
Loading apps from bootlocal isn't a good idea, because they won't have a menu button. Loading from aterm (via a bash script, etc.) will cause intermittant crashes if you're using jwm. By far, the fastest and safest way is to load from a mydsl folder at bootup, or from the mydsl panel if you need something on the fly.
I'll join the consensus regarding eliminating Firefox to save space. Dillo serves the purpose until a user gets around to understanding how to load an extention. First, however, he complains bitterly about not knowing how to update the old one.
Posted by andrewb on April 16 2007,06:30
| Quote (jpeters @ April 15 2007,14:38) |
| By far, the fastest and safest way is to load from a mydsl folder at bootup, |
& the mydsl folder can be nominated at boot time - even from a live-cd boot so you can boot a live cd with different apps loaded during the boot process, depending on what you require at the time.
Posted by MakodFilu on April 16 2007,14:53
I suggested the removal of Firefox, but I concede I *actually* use that version, not the updated one, for the sites that require me to use Javascript or something.A desktop icon stating that it will download xxMB and *then* launch Firefox would be desirable, but I don't know if that means downloading GTK 2.0 also, if GTK 2.0 would be included in DSL and GTK 1.2 left behind, if the possibility to download the outdated and hopefully small version persist or if all options would be presented after clicking.
Posted by Juanito on April 19 2007,10:40
I can't actually remember what I voted for, but the more I think about this, the more I believe it would be better to update the DSL kernel to 2.4.34 and the DSL-N kernel to 2.6.20 whilst (hopefully) keeping compatibility for the majority of old machines and maintaining a minimum set of applications.In order to keep the size down, non-essential applications could always be stripped out and posted in the repository to be loaded as extensions.
The more I compile applications/modules, the more I come across the same problem in that once compiled/loaded, the application/module will not work because the kernel needs patching. Using a patched kernel with DSL/DSL-N would be a pain, but the patches don't start until 2.4.26+ and 2.6.12+ anyway...
In summary there will usually be somebody who can post a "heroic" compilation as an extension (thus no need to concentrate on the applications) but if the kernel will not work for that application then you face a show-stopper (thus better to concentrate on upgrading the kernel).
Posted by roberts on April 19 2007,16:00
I think that this will be the approach taken.First a 2.4.34 upgrade then ...
Some apps upgrade with libs etc.
This will be a phase in approach over several/many future releases.
Posted by ^thehatsrule^ on April 19 2007,17:59
I haven't voted yet either (still thinking), but I would probably like a 2.4 upgrade first. As for the libs, that is welcome as well (particularly the libc.. but that may cause a decrease with woody repository compatibility?). And the tiny core would be nice to have, but I don't see that coming.Since 2.4.31 was reverted to 2.4.26 due to incompatibility with some old hardware, will a 2.4.26 version also be offered in addition to 2.4.34?
Personally, I am running a live version that contains both, but the size is considerably large with the added modules (and this doesn't even count the additional modules that DSL contains for other devices).
One of my ideas (reminder: this is just me testing things out) is to separate the modules from the rest. For example, KNOPPIX/KNOPPIX and a KNOPPIX/MODULES. That's okay so far changes are transparent to the user. Where I can see this come into play is the ease of switching kernels+modules, and toram. Right now toram will load KNOPPIX to memory, but this way you won't have the uneeded modules also stored in memory (saving 8-16 mb?) and they're mounted in MODULES loaded.
This method could _possibly_ open up for toram use in lower-end systems. Of course, there's probably many problems and questions with this that I can't see right now. Like if someone wants to eject the media, is it safe to unmount MODULES? (probably have to modify modprobe, insmod etc. then to ask for the media). This is possibly off-topic so I'll back off now :P
Posted by roberts on April 19 2007,19:17
I am in favor of backporting and maintaining DSL 2.4.26, whether it will be called 3.4 or 3.3.1. In fact I plan on another 2.4.26 release before releasing anything else. As long as there is interest in DSL classic I will maintain it. I have learned from the 2.4.31/dsl 2.x.adventure to not do otherwise.
I also have a PXE version of DSL that I will likely post as an additional image. Many have been asking for this. This will mean that I have 5 versions (images) of DSL.
You are right about breaking Debian Woody with these phased in updates. That is really the big question. Is the effort worth it, If it means no Debian Woody compatibility? These are issues that John and I are discussing and why I posted this survey. It seems to me, that one of the big factors in our decision on 2.4kernel version is how important is Woody.
I could also start anew with Debian Etch but that would likely be a much larger system. Trying to maintain current Debian compatibility seems to necessarily imply more bloat, which is contrary to the original design of DSL
And then there is DSL-N which I have already stripped to a tiny core that has more current libs and could be updated to the latest 2.6.20. I just recently updated Lua and all the Lua scripts.
Posted by Felson on April 19 2007,21:40
ok... I never read all 12 pages of comments, so feel free to flame me... Here is my take on all this. I think that the 2.4 kernel is fine. 2.6 is bigger, and I am not 100% sure what we would gain from it. I also think that removing any programs that are not needed in the base is a good idea. Just keep the stuff that absolutly needs to be in there. We can make extentions for anything else. The way extentions load on boot is nice, and mostly transparent anyway.
Now for my opinion on Woody... The stuff you can install from apt is almost 2 years out of date. Building stuff from source really isn't that hard. I made a dev extention with my gtk2 stuff, so really, that is the hard stuff anyway. GAIM and most other apps are cake. I think at this point if we are going to have anything that is even remotly up to date, it needs to be built from source anyway. If the extentions are all 2 year old programs, people won't want to use DSL execpt for verry special applications. Every extension I have made useing the Woody stuff has resulted in allot of bugs, or huge dsl/unc files. Ya, apt was a nice invention, but isn't that really what dsl/uci/unc is anyway?
Posted by WDef on April 19 2007,21:47
Glad there's no move afoot to lose the heart of dsl - a mean little 50MBs for which stability, completeness and lightness are the dominant requirements. I like new packages where these make a big difference eg media players, encryption etc. But for the vast majority of basic CLI apps, it's not always easy to see what you get by upgrading to bigger and fatter.Lately I've been looking at Puppy and I'm struck by just how different the two approaches really are. Different. Puppy has the whole modular, scalable thing really worked out, though the resulting diversity of releases (many by community members, with Barry's blessing) is initially confusing. It takes a while to work out what you are supposed to be downloading! But the lesson there is that the diversity has tended to strengthen the community, rather than weaken it (the forums are very strong), and has attracted very committed developers ready to take over when Barry reduces his role (apparently planned to be at the end of the year). And their repos are a bit of a mess - there's unofficial packages all over the place and it;s hard tok knwo what works and what doesn't. DSL's centralized repo with testing is to be applauded.
John Murga produces a small Puppy, but I don't think there's any doubt DSL remains the No1 50MB ruling distro. Once you get bigger though, Puppy has a lot going for it.
DSL is simpler and cleaner in many ways, but without Puppy's gui niceties (they have some nice - if perhaps verbose - guis, which are very appealing to newbies but of passing interest only to geeks.
The simple purity of DSL's uci/unc system also really shines - the equivalent in Puppy have to be loaded during boot, and can interact buggily with their regular packages (so I discovered to my chagrin). But I like Puppy too.
A scalable DSL-N could give Puppy a run for its money though in the 2.6.xx area.
Anyway I thought some of these comparisons might be helpful.
Posted by curaga on April 20 2007,16:32
I have an idea for that KNOPPIX/MODULES approach. Why not mount the modules dir by supermount? It's a readonly cd, and supermount umounts stuff automagically when the media is removed, and if writing is in action, it won't let you eject the media (though that's not possible with floppies).Then if one tries to modprobe/insmod something, it tells module isn't found, and this is the only part we need to change. It's a lot easier changing only a print line, not touching the source. It could be changed to "Module <mod> not found, you sure the cd is still in the drive?"
Posted by roberts on April 20 2007,17:03
I already have the separation of the modules via the mydsl directory and autoloading during boot, e.g. mydsl/modulesCopy a module(s) into this subdirectory off of the mydsl directory and your modules seamless load. Works just like mydsl extensions.
I had tried to remove the least often used modules, scsi and netfilter, during the last RC cycle, only to have complaints. I believe DSL is the only small distro to carry so many modules in core. I could easily remove > 2MB without affecting the majority of users.
Posted by roberts on April 20 2007,17:18
As for the question of Debain Woody and staying compatible, someone mentioned just to compile new(er) applications. That Debian compatibiity is no longer that important. But I must say to only look at DSL-N which was released without Debian compatibilty and that turned into a major gripe!Personally, I don't think Debian compatibilty is necessary in a tiny distribution. Especially one that strives to be nomadic and thereby carry the least amount of code,docs, etc.
DSL has never really been Debian compatible as one cannot do an apt-get upgrade without breaking DSL's X server.
I have always maintained that Debian's packages are not fine grained enough for me. When I ask for one package and many other packages need to be loaded when that is not necessarily the case, nor my desire. I understand that expediency most often reduces choice. Whereas DSL thrives on choice. Perhaps some would say too many choices.
Posted by WDef on April 20 2007,21:28
| Quote |
| When I ask for one package and many other packages need to be loaded when that is not necessarily the case, nor my desire. |
I think the Red Hat world is worse in that area. You want to install one thing and find there are a million dep rpms that download first (if you can find them - some fedora deps for some apps are scattered across repos). Debian package management is simpler, easier and better organized. And Yum sucks bigtime.
Posted by hrwusesdsl on May 05 2007,04:23
Robert, the view you obtain from the forum is very warped. People who take the time, and have the nerve, to post, are far more computer savvy than the real public. I will bet money there is a large "silent majority" out there who will not vote here. What that silent majority needs is friendly, friendly, friendly.And Internet.
Forget Debian compatibility. Forget most Linux experts. Go for the masses. That is your greatest service.
Posted by jpeters on May 05 2007,06:27
| Quote (hrwusesdsl @ May 05 2007,00:23) |
| Forget Debian compatibility. Forget most Linux experts. Go for the masses. That is your greatest service. |
The "masses" are stuck with windows, bloat, ads, spyware, and viruses on expensive machines. That's just the price of ignorance.
Posted by stoneguy on May 06 2007,15:32
We all choose what to be ignorant of. I can't fix my car, so I take it to the dealer for maintenance. I go a Dr. when I'm sick because I don't know enough to diagnose myself, and the laws won't let me write myself a prescription anyway. The traditional Linux attitude is "If you're not willing to pay a knowledge cost, you're stuck with Microsoft." I find that truly arrogant. I'd like to see this replaced with "You don't have to pay the monopolist OR struggle to learn something arcane. Here's something that's easy and free. Enjoy."
Choosing to skill yourself at computer usage is not inherently better than choosing plumbing or carpentry.
Posted by hrwusesdsl on May 06 2007,19:01
Right on Stoneguy. And it must be noted that the pioneering leaders in friendly Linux are undoubtedly DSL; Knoppix; Ubuntu; Puppy; and SLAX. All I'm saying to Robert is: Don't sacrifice any friendliness. I voted for tiny 2.6 core with most apps in myDSL based on my guess regarding what is most friendly / least intimidating to newbies.
Posted by ^thehatsrule^ on May 06 2007,23:28
About debian/apt compatibility: this is great for new users, esp. those on hd-installs, that can easily try to install software without asking for a mydsl extension or learning how to compile from source. If it isn't in mydsl, I try apt next. Even though there are many incompatibilities as it is, it's still great when it works - saves you a lot of time. I don't think we would ever get a true account of which features of DSL are actually used the most, etc. but I actually think that most new users to DSL try apt for their fav. software.| Quote |
| The traditional Linux attitude is "If you're not willing to pay a knowledge cost, you're stuck with Microsoft." I find that truly arrogant. I'd like to see this replaced with "You don't have to pay the monopolist OR struggle to learn something arcane. Here's something that's easy and free. Enjoy." |
Posted by buzzard on May 07 2007,01:44
| Quote (hrwusesdsl @ May 06 2007,15:01) |
| Right on Stoneguy. And it must be noted that the pioneering leaders in friendly Linux are undoubtedly DSL; Knoppix; Ubuntu; Puppy; and SLAX. All I'm saying to Robert is: Don't sacrifice any friendliness. I voted for tiny 2.6 core with most apps in myDSL based on my guess regarding what is most friendly / least intimidating to newbies. |
DSL certainly is friendlier than ZipSlack, which is what I had tried before that. In zipslack, if you start x without realizing you have to configure it first it tries to fry your monitor. DSL boots good on most PC's up to desktop, with icons already there, and a help file showing on the screen in Dillo.
Another thing I like is that, being knoppix based, you can hack the crap out of it without fear of screwing it up, because as long as you keep a backup copy of backup.tar.gz you can get it right back.
Posted by buzzard on May 07 2007,01:48
I downloaded a copy of DOSEMU from, uh, somewhere..? I forget where, but it wasn't from the DSL site. Anyway, it runs good, and way faster than dosbox, although some dos apps aren't getting the sound support (quake 1 gets it, jill of the jungle doesn't). I think that'd be a cool thing to put in a .uci file.
Posted by clivesay on May 07 2007,02:12
Some may be misunderstanding. The reason that debian compatibility would be good is FOR user-friendliness!
If DSL maintains compatibility with a branch of debian then a much greater number of community members can help build the mydsl repositories.
Newer people may not be aware of the deb2dsl or dsl2unc scripts. These are scripts submitted by community members that provide a great service! With deb2dsl, a user can install an application via apt. If they like the app and it runs well, they can run deb2dsl from a commandline, answer a few questions, and now they have a mydsl extension of the application! Want to run the app on a lowram machine using unionfs? Run the dsl2unc script on that mydsl app and now you have it as a unc!
THAT, my friends, is why maintaining some type of debian compatibility would be good. It's not really for the hd-installers, it's for the community to help build the mydsl repos to support the frugal/live cd/usb installations.
Posted by mikshaw on May 07 2007,03:04
Funny how this "user friendly" argument never ceases. Also funny how I always feel compelled to get mouthy about my position on the subject....There's nothing arcane about Linux. It's quite the opposite. Everything about Linux and GNU is wide open, and freely available documentation and other forms of help are spread all over the internet. I might say that Windows is more arcane, as it is completely closed along with a majority of the software and hardware made for it.
As ^thehatsrule^ said, no operating system is going to be easy to use for someone new to it. The seemingly simple concept of clicking stuff isn't even terribly obvious to someone who has never done it before.
The concept of "friendly" is subject to the individual user's frame of reference and personal taste, and cannot be boxed into a single collection of features. A user who is quite comfortable with a Windows interface and its tools is obviously going to have to do at least some learning and adjusting when moving to any non-Windows system. The more unlike Windows that system is, the more learning the user needs to do.
This does not mean that the other system is less friendly than Windows; it means it is *different*. Linux is an extremely powerful and flexible system, much more so than anything ever released by Microsoft. This is a simple fact. In order to harness this power and flexibility, you *must* learn more than how to click. And this clicking is what the Windows user always seems to see as "friendly". It has to be clickable or it's somehow ancient or too complicated. But a graphical interface using current technology can do only so much before it takes more effort to use than what it was designed to save. The more functionality is given to the gui, the larger, slower, and more complicated it becomes. More complexity means it becomes harder to learn and is more prone to programming errors.
The reason I spent so much time talking about Windows and its graphical interface is because every time there is a debate on user friendliness, it basically comes down to people coming from Windows and seeing that things are done differently. For these people I would say that they should consider trying a different Linux distribution. There are several that make the transition from Windows much easier. DSL cannot be terribly friendly in this sense of the word because it is targeted toward old hardware that may not have the space and processing power to handle complex graphical applications. Roberts has put a lot of work into creating a collection of small, simple tools which certainly help in this area. It's actually quite surprising what can be done in DSL without even needing to open a terminal.
For my own idea of friendly, I prefer simplicity and the flexibility and power to do what I want with my system. This is pretty close to what DSL is. If DSL were focused mainly on binary graphical applications, I wouldn't even bother with it.
Posted by jpeters on May 07 2007,06:29
| Quote (clivesay @ May 06 2007,22:12) |
| Newer people may not be aware of the deb2dsl or dsl2unc scripts. T |
It would be nice to see the wiki updated to show how to include icons.
Posted by mikshaw on May 07 2007,14:08
I think the "create tar.gz from source" touches on that subject. I think you're right, though, that it should be seen in other sections as well. Perhaps a generic "basics of building myDSL extensions" should be written to explain all things that are common to all mydsl types: icon, menu item, declobbering, submission, etc.
Posted by hrwusesdsl on May 07 2007,15:43
I forgot Linspire / Freespire in my list of ground-breaking, truly friendly Linux distros.Mikshaw, what do you think makes DS less friendly than Puppy? It is, in case you don't know. Damn Small has advantages in things like the really slick encrypted CONFIG file, truly great for Internet access. I use both distros.
Posted by MakodFilu on May 07 2007,17:01
DSL *is* "user friendly".<rant>
What DSL *is not* is "Windows user friendly", and I like it that way. However, there are some resemblances, enough to make a Windows user think it is viewing something familiar, but that is it, just a bonus, something to be glad at, not the main goal. Besides, Windows interface plainly sucks: the whole idea of floating windows you have to move and resize almost every time is sick and ill-minded. Folders in the desktop is even worse, you have to minimize every window just because there is not a "home directory" culture and way of thinking.
</rant>
Posted by jpeters on May 07 2007,18:13
| Quote (mikshaw @ May 07 2007,10:08) |
| I think the "create tar.gz from source" touches on that subject. |
Thanks, Mikshaw. I hadn't seen that (although I'm not sure it addresses how icons copied to .xtdesktop get removed on reboot).
Posted by mikshaw on May 07 2007,18:42
| Quote |
| I forgot Linspire / Freespire in my list of ground-breaking, truly friendly Linux distros. Mikshaw, what do you think makes DS less friendly than Puppy? It is, in case you don't know. |
Again, as so many people have repeated, "friendly" is subjective. What you are describing is, as MakodFilu just put it, "Windows user friendly". Linspire was created specifically for users who want the Windows version of friendly. You obviously fall into that category. Nothing wrong with that in itself, but I'm just trying to get you to understand that Linux is not, and never will be, Windows, and many Linux developers (and users) have a very different view of what defines friendly. It's not being elitist or arrogant to suggest staying with Windows, or choosing a distro that is more like Windows, if that's what you like...it's just being logical. I came from Windows myself, and of course it was a rather steep climb over the first few hills. I did not, however, expect it to behave like Windows. In fact, I was surprised that the most popular distros were much more like Windows than expected.
I don't know much about Puppy, apart from a very short test of Grafpup, so I have no opinion of that.
EDIT: there is one thing mentioned a lot by the "user friendly" people, and that is that Windows and Mac are both designed with the concept of "just getting your work done and not tinkering around". I can't argue with that, and I can't deny that most people who use Linux regularly seem to have at least a little bit of tinkerer in them. Personally I love to tinker, and I honestly might still be using Windows if I didn't enjoy messing around with the operating system. After becoming used to Linux, though, I wouldn't ever go back even if I suddenly lost the tinker gene. Just having the freedom from spyware, constant virus checks, pushy software developers, and instability is enough to keep me here.
Posted by jpeters on May 08 2007,03:14
Maybe we could change the topic back to what works best, and forget the "friendly to the masses" thing all together. Trying to please the masses only means one thing: lowering the quality. It's that way in art, music, theater, literature, journalism, science, politics, and even religion. There's only one reason to even attempt to please the "masses," and that's to make money. Now if there was, in fact, any money in DSL, Robert would be rich by now and the DSL store would be in the process of being acquired by Walmarts.
Posted by lucky13 on May 09 2007,22:42
mikshaw:| Quote |
| I don't know much about Puppy, apart from a very short test of Grafpup, so I have no opinion of that. |
I've played around with Puppy a little more the last couple days. There are some things about it I find commendable (such as wizards for everything and unified dialogs for various things that should be grouped together) and some things inexcusable -- and I suspect those things are done at the expense of making it easier.
Such as:
- Puppy runs as user root. Win95 users will feel right at home with it.
- Puppy loads to RAM by default.
- Even though it's a "small" distro, it's not oriented to staying light (look at the community offerings of KDE, etc.).
Etc.
Also, about your edit about Macs and Windows: I don't think either is particularly maintenance-free. I "tinkered" with more registries, ini files, etc., than I ever wanted to -- and that was out of necessity, not just the sake of tinkering. Macs aren't always user-friendly or tinker-free, either:
< http://www.macfixitforums.com/ >
jpeters:
| Quote |
| Maybe we could change the topic back to what works best, and forget the "friendly to the masses" thing all together. Trying to please the masses only means one thing: lowering the quality. |
I don't think the two are mutually exclusive -- I think they actually go hand-in-hand. New users come with the territory when things work better, more easily, and more efficiently. Innovation and development reduce the learning curve for others who are hindered by certain levels of technology. I don't find that something to avoid, but something that should be embraced.
What makes things easier for "the masses" also improves things even for those who are comfortable in the console. If nothing else, the masses buying computers and peripherals and using the Internet at the rate they have over the last 15 years has spurred the industry so we're no longer paying as much (or more) for slow computers than new cars cost and we're using CDs, DVDs, and USB sticks instead of floppies and even Zip drives. I don't think quality has suffered one bit at the expense of user-friendliness. If anything, it's improved greatly.
Posted by jpeters on May 10 2007,00:59
..and it gets even better if you don't use windows (faster, cleaner, cheaper).Note: I also own a windows computer for specific functions (recording studio, etc), mostly because I haven't found the need to go pure linux.
I've also spent endless hours going through registries and reconfiguring windows systems that suddenly stop working. With DSL, you just boot up another copy......can't get much easier than that.
Posted by hugglebear on May 23 2007,10:06
I believe DSL to be the future of computing!In the future we won't need to carry laptops etc, we will just plug in to a dumb terminal and go. DSL provides this environment already..... Almost!
THREE IDEAS
1. Make it work!
If a version of DSL can be created that works on EVERY computer, then the job is done. I have problems with my sound card and printer which I will address in other posts, but these are problems I simply don't care to have. When an idiot like me powers up his PC, he just wants it to freakin' work. He doesn't care why, where or how, and will go to whoever can provide this.
I would imagine that the way forward could be accomplished by simply leaving out ALL applications that are not requisite to boot/install DSL. Why waste space including editors, graphic programs, office apps etc... when they could all be downloaded via MyDSL. If you need it, then download it!
For me the flexibility of the MyDSL feature, is one of DSL's strongest points, and could be more fully realized.
2. Let the DSL community help!
DSL is already well known in Linux circles, but the market is not in preaching to the converted. If we can convince just a tiny part of MS users to switch to a fast, safe and secure environment (which shouldn't be too hard, now should it?), then DSL's future will be secure.
But how?
I would like to become the official Central European Distributor of DSL (unless of course you have one already). I believe in DSL, and I am quite prepared to stand for hours at a car boot sale, market and Internet cafe's etc... selling it, I believe others are too. So let me/us help finance your fantastic work by setting up a program that allows official distributors a bulk price on the DSL Biz Card CD. This would allow the DSL community to make a small profit from each sale and help finance DSL's development. The distributor should not be allowed to price his stock over that of DSL, thereby (I believe) keeping the integrity of EULA. If the licence were to be compromised by this, then perhaps letting each distributor add his own user manual with translation where required, adding value!
3. You're gonna' hate me for this!
Make DSL more Windowish. There, I said it!!! Now I don't mean that you should use your hard earned dollars developing a blue screen of death, rather make DSL FEEL like Windows without taking away the control, power and flexibility of Linux.
Windows: This is like a brand new car that doesn't work. It looks great, but what the hell use is it!
Linux: Purrs like Ferrari, but has a bitch of a learning curve if coming from Windows. (I overheard someone saying that "Life is to short for Linux")
So here comes the rocket science.... Take the good and leave the bad!
Freespire has achieved this! But it doesn't have that orgasmic Toram cheatcode making my PC work like lightning, nor the "OS in your pocket" style 50mb biz card CD. They have also set a phenomenal benchmark. Simply go to the freespire Wiki (http://wiki.freespire.org/index.php/The_Perfect_Desktop_Linux_OS)
At the bottom you'll find this challenge...
----------
"How does YOUR desktop Linux system stack up? To test it and find out, go to this test page (http://www.linspire.com/products_linspire_whatis.php?tab=compatibility) , and start clicking on each of the links. How did it do? Did it handle all the different file types and multimedia formats? Did it play everything in-line, and properly?
With Freespire, all of the links will work, without any additional software or tricky "clickety-clack" required."
----------
I think that DSL could pick up this gauntlet and smack Freespire in the face with it, but I don't know all the extensions I need to download from the repository in order to achieve this, Do you? A DSL wiki page on this subject may well be a good idea.
Of course it maybe that I'm riding the wrong horse, and should suggest to Freespire to include the magical features of DSL...!
But I'm here, and I am ready to do battle to help DSL become the most popular/ easy to use Linux distribution, and who knows, perhaps we can all make a few bucks along the way.
What say Ye?
Posted by lucky13 on May 23 2007,15:07
Linspire fanboy wrote:| Quote |
| 1. Make it work! |
It does. The problem with making it work on "EVERY computer" comes down to hardware issues, not software issues. You should direct this point to hardware manufacturers who refuse to open their drivers so people who buy their products aren't locked in to one (Windows) or two (Mac) particular platforms. Better yet, don't buy stuff that doesn't play nice with whatever OS(es) you choose to run. And I'll go a step further and say, Don't support distros like Mint and Linspire that support closed hardware because it only encourages vendors to sell closed hardware.
| Quote |
| 2. Let the DSL community help! |
That's what this forum is all about.
| Quote |
| I am quite prepared to stand for hours at a car boot sale, market and Internet cafe's etc... selling it, I believe others are too. |
Some of us believe in DSL enough that we choose to make copies and give them away. There's nothing to stop you from trying to make money off the work of others, but I encourage you to give free copies to a few friends and ask them to do the same if you (and they) really believe in it.
| Quote |
| The distributor should not be allowed to price his stock over that of DSL, thereby (I believe) keeping the integrity of EULA. |
Here's the "EULA": < http://www.gnu.org/copyleft/gpl.html >
| Quote |
| Make DSL more Windowish. |
No. Hell no.
| Quote |
| Freespire has achieved this! |
Good for Freespire. Why should DSL do what others are already doing? Go sell Freespire out of your car instead.
| Quote |
| (re: Freespire's "challenge") How does YOUR desktop Linux system stack up? |
Well enough. Freespite/Linspire supports closed-source hardware and software. DSL can do a lot of that if the right extensions are added. I disagree with those who think a GNU-Linux distribution should include closed-source drivers and software by default; that should be left to the end user to add if he or she wants it.
| Quote |
| With Freespire, all of the links will work, without any additional software or tricky "clickety-clack" required." |
Then that's what you should peddle instead of DSL. Adding extensions isn't "clickety-clack" and I don't think it's DSL's responsibility to support closed-source code like Flash, Real Audio, Windows Media formats, Apple's AAC, etc.
| Quote |
| But I'm here, and I am ready to do battle to help DSL become the most popular/ easy to use Linux distribution, and who knows, perhaps we can all make a few bucks along the way. |
"Do battle"? Other than asking the developers to bloat the code with proprietary codecs (so DSL can play certain media files and web content like Flash out of the box), just how do you intend to "help DSL" become more popular? By selling it? When it's already free?
| Quote |
| I think that DSL could pick up this gauntlet and smack Freespire in the face with it... |
Did you read system requirements for Lin/Freespire?
# PC with 800 MHz or higher processor
# 128 MB of RAM (256 MB or higher recommended for best performance)
# Hard drive with 4 GB free space
And there's more they recommend!
< http://www.linspire.com/products_linspire_whatis.php?tab=systemreq >
In comparison, DSL can run on a 486. DSL can run on 16MB RAM, and can be loaded to RAM on systems with 256+RAM. DSL can be run off a CD, off a USB thumbdrive, or embedded in another OS. It doesn't require 4GB on a hard drive, either -- try 50MB, plus a reasonable swap partition, and however much a user wants to add for /home. Even a traditional Debian-style installation of DSL only needs ~200MB, far less than Linspire requires.
Face it: DSL kicks Freespire's fat, bloated ass. Users shouldn't be locked out of using Linux because it requires 800+Mhz chips, 256MB RAM, and at least 4GB space on a hard drive. I'm for making things easier for all users, but Freespire should aspire to DSL's standards on running on smaller and older systems and not the other way around. There's no such thing as an obsolete computer -- there are just bloated operating systems. DSL shouldn't become yet another one.
Posted by mikshaw on May 23 2007,15:40
1. Your wish for it to "just work" on all hardware is a common one. However, unlike Mac, Windows, SGI, or whatever other machines that are designed for a specific operating system, Linux doesn't have hardware manufacturers catering to it, so many times developers must use a trial-and-error approach to writing drivers.Secondly, DSL has a 50mb size limit. I'm not sure it would be possible to include drivers and boot-time checks for all available hardware. Basically, there is no way you can please everyone....something has to be left out. Besides, even Windows doesn't always "just work" when it comes to hardware. I've had to manually install more drivers for Windows than for Linux over the years, because Windows tends to drop support for older hardware when they release new versions.
Third, how come no one wants to bother learning something about their operating system anymore? Just about everyone I know understands how their cars work, and can fix minor problems, but when it comes to computers they are completely clueless. One of the main things to consider when installing Linux is what hardware you have. Do a little research; find out whether your hardware is supported by open source drivers. If you have a device that is not supported, or is poorly supported, spending a few bucks on something else will save you a lot of trouble. What it comes down to is that the Linux community is not at fault if your hardware doesn't work. Many hardware manufactures keep their specs a highly guarded secret.
2. Linux, including DSL, is an open-souce community project. This means you already have the permission to distribute it. You don't need permission, because it is already granted.
However, you MUST read and understand the General Public License before you begin distributing DSL. Some of your comments lead me to believe that you do not understand the concept of software freedom as defined by the GPL and the Free Software Foundation. The GPL is the deciding authority of how DSL can be distributed, not the DSL developers (apart from possible trademarks). As far as I know, everything in DSL is comes under GPL-compatible license. Also, please refrain from the use of the term "EULA" when referring to open source projects...it will piss off a lot of people.
3. For pete's sake, NO!
This topic regularly comes up in every Linux forum all over the world, and it is usually (always?) posted by someone who is accustomed to using Windows. The response is usually an emphatic "Linux Is Not Windows":
< http://linux.oneandoneis2.org/LNW.htm >
Freespire (Linspire) was developed to look and feel like Windows. DSL was not. If you want the look and feel of Windows, use Windows or a clone of Windows. Otherwise, please learn the strengths of Linux.
Posted by jpeters on May 24 2007,08:10
| Quote (mikshaw @ May 23 2007,11:40) |
| If you want the look and feel of Windows, use Windows or a clone of Windows. Otherwise, please learn the strengths of Linux. |
I just spent about 2 days trying to get a Windows XP computer working again. First, you need about 4 different programs to keep removing adware, spyware, worms, various viruses, etc., in addition to your firewalls, etc. Also, a registry cleaner is required. All these programs need to be constantly updated, and eventually start doing more damage to the OS than what they were initially intended to prevent. Next, the uninstallers frequently don't work, requiring the user to manually go into safe mode, and then prune the registry after deleting files in the directory. One damaged systems file somewhere, and the OS starts crashing on boot, sometimes requiring reinstalling the OS and starting all over (which can take hours).
The most I've had to do with DSL is boot a backup disk. I suppose if I really wanted the look and feel of Windows, I could throw the computer down the stairs, jump on it few times, and hopefully get the thing to keep crashing and generating blue screens. It would also be important to have a little hour glass go off every time you click on something.
Posted by hugglebear on May 26 2007,10:43
Respect is something both earned and deserved, and when meeting for the first time, I give everyone 100% right of the bat. From then on, it rests on the individual concerned. Of course the Internet is the great equalizer, yet common courtesy is still high on my list of personal values, just as my mother taught me.I took a good deal of time and effort in writing and re-writing my previous post so as to present my ideas with clarity and fairness, in response to "Moving Forward". I may well have failed to do that, but I find no place in such a forum topic for insults and sarcasm, I am no more a Linspire fanboy, than you are Bill Gates' Bum Boy! I feel like a politician that was deliberately misquoted or misunderstood, but those are my insecurities and shall remain my own as they equally have no place here on this thread.
So if I wasn't clear before.... I LIKE LINUX, I LOVE DSL!!!
I could download the current.iso, order a bunch of business card Cd's and setup my own little production line right here in Berlin, Germany. And nobody would know the difference, and even if you did, exactly how much money do you have in order to take me to court.
You see my point!
That is of course NOT what I want to do. What I want is to order from the DSL website multiples of let's say 100 Biz card Cd's, but at a reduced price. This will reduce profit, but increase turnover. Somehow I don't really think I need to explain this to you, you are, simply because of your presence here, intelligent people.
Who was it that said...
"If I met bill gates, we'd have nothing to say to each other. He couldn't teach me anything about computers, and I couldn't teach him anything about marketing!"
My post was based on the $1 challenge. I don't have a job and therefore have no money to simply give away. But still I would like to compensate the developers of DSL the best way I know how. And a little intelligent marketing could go a long, long way to helping Robert and the rest of the team in working full time. "And then imagine what could happen to our little Linux project!"
Robert! By way of showing my personal commitment, I will translate into German all texts sent to me, so that a German version of dsl can be released in the future.
I have no more to give... Take it or leave it!
Posted by ^thehatsrule^ on May 26 2007,20:07
Afaik gpl'd (v2) code can be (re)distributed any way you like... so I don't think you read the license properly?About making things look like windows: they probably tried to copy others... (i.e. the app/taskbar on the bottom vs the top (macos) just to avoid problems, but it is ergonomically better to have it on the top... I set my *box this way heh...)
Anyways, I think this is getting off-topic.
Posted by roberts on May 27 2007,19:37
| Quote (hugglebear @ May 26 2007,03:43) |
| ... What I want is to order from the DSL website multiples of let's say 100 Biz card Cd's, but at a reduced price. This will reduce profit, but increase turnover. ... Robert! By way of showing my personal commitment, I will translate into German all texts sent to me, so that a German version of dsl can be released in the future. I have no more to give... Take it or leave it! |
John and his Company called Lizard Biscuit handle the store. I do not have anything to do with hardware items available in the DSL Store. That is a separate operation. You would need to contact John for anything regarding the store.
Have you seen this:
< http://www.damnsmalllinux.org/index_de.html >
If you find additions or corrections please email to us.
Posted by mr sparks on May 27 2007,22:29
If we need to support both new apps and old hardware but stay within size constraints , it sounds like a serious choice.If the future is with the new , a purpose configured OLD version for old hardware ,needs to be identified as such.
with a large sticker/Flag --Old Sytems Here--
and its hardware abilities sumarized and documented .
(Easy-USB-internet)
(Save files to USB-stick)
As to a choice so long as it works and I can find what I need,I'll leave it to the experts.
Thanks for all the great tools
Posted by sankarv on June 01 2007,04:41
it would be better to provide additional support for hard ware or applications in the form of extra mydsl apps so that DSL can be used for both old/new hard wares and also the basic goals of DSL are not affected.
Posted by daldred on June 06 2007,19:35
Just found this debate...My perspective: DSL is the only distro I've managed to get loaded onto my tiny machine - an old Libretto with 32Mb of memory.
Now that's a seriously low-end machine these days (P75, 32Mb of RAM, HD upgraded to 6Gb) - but DSL runs on it quite nicely, does wifi and connects me to the net. Browsing with Dillo is reasonably fast; even with Firefox it's not too bad, once Firefox has actually launched!
To me, the 50Mb limit matters much less than the way it will run on a low-mem machine. Space isn't really my issue!
Would it be feasible to have DSL 'cores' set up for low-spec / high(er) spec machines, then build everything else round them in modules?
Posted by mikshaw on June 06 2007,21:28
| Quote |
| P75, 32Mb of RAM |
| Quote |
| To me, the 50Mb limit matters much less than the way it will run on a low-mem machine. |
Posted by Zucca on June 06 2007,21:49
"Forget the Kernel update the Apps in DSL"
Posted by daldred on June 07 2007,16:43
| Quote (mikshaw @ June 06 2007,17:28) | ||||
|
| Quote | ||
|
I think the Libretto originally had 16Mb, upgradeable to 32Mb - but no more than that!
| Quote | ||
|
Perhaps if the 'core' of DSL (enough to get Fluxbox with networking reliably up & running on a reasonable range of hardware) were kept within a small downloadable unit (or two units, one for 32Mb or less (just SL, perhaps - too small to give a dam!), one for the big stuff), and everything else were separately downloadable as apps, you could have small downloads and both old machines and new ones could be catered for.
And there could still be a standard DSL 50Mb mini-CD as there is at present, comprising the bigger-machine version and some apps - awkward people with archaic machines which need DSL to have *any* sort of Linux would be quite willing to take the extra steps to download the 'small' core and add their own selection of apps. Our old machines tend not to like CDs anyway :-)
Posted by Key on June 08 2007,17:27
Update DSL Kernel to 2.4.34Or to whatever.
But please update the Kernel.
Otherwise I can not access the internet anymore with my new computer.
Thank you
Posted by roberts on June 09 2007,21:52
And with that I will close this topic.After 94 replies and good discussion, thanks to all!
Based on your voting and discussions with John,
DSL v3.3 will be the last in the 2.4.26 kernel series.
I will likely post sub-releases, e.g., 3.3.1 while we transition to using 2.4.34.
I will make available my kernel development tarball, so as to faciliate those who wish to contribute modules.
-- Robert