DSL revival?

[News]

News

[Stats]

Stats

[betatest3]

Will DSL, ever be revived?

[CNK]

Well John hasn't replied, so things don't look all that hopeful right now.

If you look at his profile (http://damnsmalllinux.org/forums/index.php?action=profile;u=1), it shows that he hasn't posted since January (also, the download site hasn't had a new upload since 2012), but he's marked as last active this month, so at least we're not forgotten.

For basic DSL, I can only see significant changes coming from a movement towards more custom software and scripts. There aren't that many resource conservative programs still being developed (Dillo being the main one that comes to mind) and when I compile any fancy new software, it usually isn't long before I trace a dependancy trail that leads to needing at least a v2.6 kernel, sometimes even v3. There are still a few things that could be fixed up in DSL, but I can't see where major feature changes could be worked in without harming compatibility and size.

I have to wonder how many people are still using DSL to make their old PCs useful again. Or whether many are putting it on far more modern machines just for the exceptional speed at which it can run (I have done this myself on a 1Ghz IBM Thinkpad with 512MB RAM).

It's that job which I believe was John's target application of the new DSL he said he was working on last year. It basically sounded like a new DSL-N. One can only assume that has been going slowly at best though. Yet there's nothing to stop anyone else making their own branch of DSL at any point in time.

Right now though, I'm typing at a Pentium 1 120MHz, with 81MB RAM. I wonder how many others still do? I might try to put up a poll later to find out...

[betatest3]

How much work would it take to use kernel 3.12 in DSL?

[CNK]

Well assuming you base it on Knoppix like the current DSL, it would just be a matter of slowly stripping parts out, tweaking others, and compiling updated versions of software.

Never having tried to create or significantly modify a distro, I couldn't comment very well on the tecnical complexity of the task, but it would be tedious, and little from the current DSL could really be used directly.

Not that DSL with a v3 kernel would be a valid new version of DSL, as system requirements and total size would contradict the goals of DSL. Instead this would effectively be a new DSL-N.


As for DSL itself, I believe it could still go further. While a lot of the software is no longer updated or has suffered bloat, there are still improvements that could be made in the scripts and the user interface. With enough support, it would be great to have basic software applications developed for DSL, effectively turning it into a base for low resource software development. The RaspberryPi, even running DSL through Qemu (which I'd love to try if I had one), could be a great modern application of DSL. But all this would need willing developers, I don't have the time or experiance, John seems on and off at best, and the other past contributers have gone.

[Benny]

For my part, what I see as the most urgent upgrade to be done is with respect to the browser (Bon Echo in this case) because, probably, is the most used application whenever a person turns on a computer. In my case I've a lot of difficulty or am unable to read my yahoo emails from my current version, which traditionally, has been the first task to be done; then, there comes the rest...

[CNK]

Yeah, although I often manage to avoid bloated websites (or just force them to be lean by running them in Dillo), Ebay in particular has always caused my trouble in Bon Echo (only getting worse as time marches on). Today I also noticed that one website I occasionally visit has started forcing users over to https with a security protocol that Bon Echo doesn't support, so now I can't load it at all.

Then there's Java, recent versions do work on DSL (well last time I checked anyway), but don't support Bon Echo, so they're only good for external programs that run with Java.

But the problem is that after Firefox 2 (Bon Echo), Firefox started relying on dependancies that aren't compatible with the DSL system (I don't know the details, but I'd guess the kernel was too old for something). Though I'd be happy with Firefox 3.6, I expect the latest releases would be about as compatible as a new Windows 7 program is with Windows 95 (from a non-technical perspective).

As I said before though, this isn't really what DSL is designed for. With the Pentium 1 I'm using at the moment, Bon Echo takes about as long to start up as DSL does, it's slow as hell and if you tried to load Ebay, it would spend about half an hour trying before the system would crash. But with Dillo I can browse simpler sites without a hassle, and as quickly as a modern machine running Firefox and loading all the junk that Dillo doesn't bother with. That's the system that DSL was really designed for, but it seems a lot of people (me included, for my laptop) are putting it on when really they want something with the DSL philosophy, but better suited to a later generation of PCs.

I'll do that poll I mentioned earlier when I get the chance, just need to work out how to class PC "generations". EDIT - Done (http://damnsmalllinux.org/forums/index.php?topic=876.0).

[james c]

The problem with running newer browsers ,etc . is because DSL is limited to GTK1 ....app selection is therefore very limited. Newest browsers that I'm aware of are located at http://www.lamarelle.org/mo-zi-lla/mozilla.php . FF 2.0.0.20 and SeaMonkey 1.1.19. Newer browser are generally GTK2.

[CNK]

Yeah but I'm not sure why at least V. 3.6 was never compiled with the GTK2 extension (gtk2-0705.dsl) installed.

I did have a look at it a fair time ago, but now I can't even remember why I gave up on it, though I had a lot of trouble trying to find a place to download the Firefox 3.6 source code (surprisingly).

[p3ngu1n0]

A new version of knoppix just came out in the last few days, and I believe it uses kernel 3.15.x. Maybe forking a previous version and trying to compile the latest longterm 2.6 kernel to work with it would be best? If we got that working, we might be going somewhere. I'm just hoping to see DSL leap to it's glory days again, it's the distro that got me into Linux in the first place.

[fatmac]

I'd like to see a modern version of DSL, with as much backwards compatibility as is possible using todays software & drivers, (i.e. wifi).
I don't know where our DSL leader (John) is with a newer version than 4.11.rc2 - is there any indication elsewhere that I may have missed?
Failing that, let us see how you would like DSL to progress?

[CNK]

John hasn't said much about 4.11 since the last RC release. My guess is that he lost the time to work on it again, but I don't really know. He still logs into the forum every month or two.

I don't know how one person (well, perhaps with the exception of a very good full-time software developer) could manage to really keep DSL progressing to support newer hardware and software without loosing much of the support for the older generations.

Even ignoring size (which would be a bit silly for a main DSL release), a modern kernel comes without much of the early hardware support that has been stripped off over the years. It wouldn't be as simple as adding it back again (which could be done simply using Kernel Modules), the driver software would have to be adapted to work with the cahnges made to the kernel. You'd have to make sure the software line up stayed light enough to run on a Pentium 1 as well, which could mean making changes to some application software to keep it working in harmony in a modern system.

Perhaps it would be better to keep the old kernel and "backport" (if that's the right term) the hardware support. Of course this would basically make DSL a branch from the Linux Kernel and over time it would probably become its own entity like BSD. Like I say, this would be great if someone very good was willing to devote their life to it. Ideally it would also inspire its own suite of low resource usage software to fill in the gaps that exist at present.

That second one is what I'd really like to happen, but I just can't see it. In reality, I expect the best that will happen with DSL (and it's not really all that bad) is that the standard version will slowly be refined at its current level (new versions of Dillo, better installation scripts etc.), while a new DSL-N will take over most of the development and target more recent computers (eg. last 10 or so years) with improved hardware and software support while retaining the DSL philisophy to the extent possible. That would cater to both camps that tend to pop up here on the forum, and I expect it's what John is already aiming for (or, at least was when he started the new forum).

[AE7XQ]

I would like to a see a new version of DSL, but before that can happen, there really needs be a substantial user base that would want it. DSL is still quite robust, despite it being a smaller than average Linux Distro. I remember when DSL and Puppy were almost tied in terms of being minimalist distros. It would seem that Puppy has become rather bloated. DSL is still small, save for what the end user adds on after installing it. Perhaps a DSL group on Facebook or Yahoo would help to drum up new interest and even new developers. Me? I'm just an amateur radio operator who is planning to use the most recently available release candidate to set up an HF/VHF/UHF station. I have already created a thread regarding that and have promised to post updates in that thread. However, that might be another thing, all of us who are here discussing what we are using DSL for and perhaps what we'd like to see DSL become. Lets be positive and show the world that DSL is more than what people can imagine. We can do this.

[AE7XQ]

Doing some thinking. Perhaps we can introduce DSL to a new type of user base. Specifically ARM devices, such as the Raspberry Pi. In truth, I don't know much regarding how ARM devices function, or what type of OS Hierarchy they require. However, I am willing to bet that it is more than just an EXT3 formatted SD card and a kernel swap. Still, I think if it was done, it might generate a little more interest in DSL.

[fatmac]

What about a DSL-NG - New Generation - even if it tops out around the 100Mb mark, for the newer machines, with up to date kernels - could this be practical?
(Or even as a 64bit DSL)

[Daniel]

How much work would it take to use kernel 3.12(http://insanityfittest.org) in DSL?