On Mon, 21 Jan 2002 [EMAIL PROTECTED] wrote: > Steve wrote: > > > > This isn't about Linux. It's about Michael (and Steven). > > He decided when porting Arachne to Linux, that he'd go > > with the latest libraries. > > More important than that: Michael did not release the source. > If we had the source, libraries would not be an issue.
Yes, but I was trying to go from where we are today rather than what "could've been." IF Arachne was open source, then we wouldn't be having this exchange. > > I think keeping BasicLinux small is a good goal, but > > when it's done by keeping it in a 5 year old time warp, > > (that's 35 computer years, remember) then obviously, > > there will be certain pieces of newer software that > > just won't work. > > Not true. BasicLinux uses recent versions of busybox, links, > nano, thttpd, icewm, xfreecell, etc. It was not difficult to > compile them for libc5. And they work just fine. libc5 may > be older than glibc2, but it's also smaller and faster. Is > that a bad thing? No, it's a good thing... until you try to install something like Word Perfect, or some other piece of closed source software that "requires kernel 2.2.x" Most people who are migrating to Linux today are doing so from a Windows experience. Telling them they can't run Word Perfect or Arachne because those programs are closed source won't fly with them. They just want stuff to work. > > If you want to run Linux kernel 1.0.1, you're perfectly free > > to do so. > > There's a big difference between 1.0.1 and 2.0.34 (the BasicLinux > default kernel). I was using the extreme to demonstrate the point. I can't imagine anyone would *actually* want to run 1.0.1. > The 2.0.3x kernels have been remarkably stable > and reliable. I know. I ran 2.0.30, 2.0.34, and 2.0.36. > The 2.2.x kernels have been more problematical I've run 2.2.14, 2.2.16, and 2.2.19. They have been remarkably stable and reliable. > and there are many reports of 2.4.x kernels having embarrassing > crashes. I'm not talking about newbie-crashes, but crashes on > 24/7 servers with experienced sysadmins. Yup. Let's attribute it to Alan Cox being out of the picture. ;-) > Many experienced Linux users are sticking with the 2.0 kernel > and development of that line continues. The latest is 2.0.39 > and I think 2.0.40 is on the way. Hmmm... interesting. Last I was aware of was the 2.0.36 I ran. I wonder if these later 2.0.xx kernels can do ipchains. > > Just don't expect every programmer in the world to make his > > programs backward compatible > > No. All I ask for is the source code. It would be nice, but it just isn't going to happen for everything. > > Steven seems to have his computer frozen ca. 1997. > > June 1998 actually. But that's just the foundation > (Slackware 3.5). And it's not frozen -- I've added > plenty of more recent stuff. > > It's my hardware that's frozen. My 486 runs my Linux just > fine. That would not be the case if I "upgraded" to a new > "feature-rich" library and a new "feature-rich" kernel and > a new "feature-rich" desktop and a new "feature-rich" browser. I guess we'll just have to disagree. I have RH 6.2 installed on my sandbox 486, and it works just fine... and runs, to human perception, every bit as fast as RH 5.1 did. > Unnecessary feature-creep (bloat) is killing old PCs. Whether > the feature-creep comes from Redmond or RedHat is irrelevant. You can't say the same for Win'98 vs WinME. For that matter, I seriously doubt either of the Redmond products would even install on a 486. I wish I'd bookmarked the story where a European Green organization gave Linux an award for being "green." Paraphrasing: Any time an up-to-date OS can be run on older hardware, that's one less computer ending up in the landfill. - Steve
