On Thu, Feb 26, 2009 at 10:00:34PM -0500, Jim McQuillan wrote: > > > Rob Owens wrote: > > On Thu, Feb 26, 2009 at 02:45:10PM -0600, Scott Balneaves wrote: > >> On Thu, Feb 26, 2009 at 12:48:15PM -0500, Rob Owens wrote: > >> > >> <snip> > >> > >>> ... On the other cards, I was given a terrible screen resolution > >>> (800x600 or less on my 18" lcd). > >> And this backs up my point, to a certain extent. > >> > >> When I started using LTSP back in late 1999, 800x600 was the "good" > >> resolution, > >> 640x480 was the "terrible one", and 1024x768 was the "OMG go out and buy a > >> $500 > >> video card and that super-special NEC Multisync monitor" "just out of your > >> price range" one. > >> > >> So, the hardware that would have been considered "good" back then, is now > >> considered "terrible", partly due to our own expectations, and partly due > >> to > >> the fact that it's getting darned near impossible to function on anything > >> LESS > >> than 1280x1024 since there's so much visual goo on the screen. > >> > >> The other point was brought up that X.org itself is fogetting it's "remote > >> display" roots, and going for the bling. > >> > >> Certainly, us LTSP developers are pretty cognizant of the lower end boxes. > >> Myself and Jim know a lot of people in Brazil who don't have access to > >> anything > >> else, vagrantc works with FreeGeek, and they have LOTS of older hardware, > >> and > >> Gadi makes a business out of selling nice, small little boxes. > >> > >> However, the people who make the bits we USE, like Xorg, the kernel, etc. > >> aren't really thinking about us anymore. The rush to finally realize "the > >> year > >> of the linux desktop" and match Windows/Mac "bling for bling" is putting > >> less > >> of a premium on nice, small, and tight code, and more on getting super-sexy > >> speedy features in quickly. > >> > >> That having been said, I still think LTSP's doing better than most. I'm > >> sitting here on my 1ghz, 256 meg workstation, which, if it were running > >> full > >> Linux would be pokey or (heaven forfend) Vista, would be unusable, and as a > >> thin client, it's fine. > >> > >> We might be able to come up with a scaled back kernel for each distro, and > >> a > >> "hand crafted" set of udev rules, but in the long run, it's a bit of a > >> losing > >> battle. > >> > > So what about my idea of utilizing the lightweight distros for the chroot? > > If somebody can give me some guidelines about how it would work, and what > > the distro maintainers would need to do, I'll start joining their mailing > > lists. I've got no job for the next 2 weeks -- I'll get right to work on > > this > > if the LTSP developers think it's a feasible idea. > > 2 weeks? That's just enough time to start thinking about it. > > To actually integrate LTSP into a distro, such as Damn Small Linux (DSL) > will take months and months. In addition, there's almost certainly > changes that will need to be made to some of the distro packages to make > them behave properly in a network-booted environment. > > What we've found is that it really helps if one of the developers from > the distro gets onboard to make this happen. That's the way it has > worked so far. > That's what I meant to do in the next 2 weeks -- check out distros and see which ones are interested in working with us. I realize that to actually produce something usable will take a long time.
If I approach the developers of the lightweight distros, at this point I don't know what to ask them. Could somebody point me to a document describing what they will need to do? In general, if I want to compare the suitability of a distro to be run as the LTSP chroot, what is a good performance gauge? Boot time to a command prompt, maybe? > You could cobble something together using some DSL bits, but if you > don't actually use their packages and package management system, it'll > be nearly impossible to keep up with security updates and bug fixes that > the distro already handles. > I'm not planning on putting together a "more modern LTSP 4.2". I'm interested in using the muekow methodology with distros that are designed to be lightweight. -Rob ------------------------------------------------------------------------------ Open Source Business Conference (OSBC), March 24-25, 2009, San Francisco, CA -OSBC tackles the biggest issue in open source: Open Sourcing the Enterprise -Strategies to boost innovation and cut costs with open source participation -Receive a $600 discount off the registration fee with the source code: SFAD http://p.sf.net/sfu/XcvMzF8H _____________________________________________________________________ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.freenode.net
