(old)Puppy Linux Discussion Forum

BIOS settings seem to be critical. That brings one huge and immediate problem - 90% of users have never seen a BIOS screen! If they did, it could bring disaster if they change the wrong setting. Worse, there are far too many brain-washed users still buying proprietary boxes instead of doing their own PC assembly. ipso facto, they are stuck with those dreadful truncated Phoenix BIOSes requiring to mount a (W98?!) hard disc to use a utility to open up the hidden settings. Activating those hidden settings can be a risk even for hardened fiddlers.......
Probably it can all be done in Linux? That reduces the field of those that can to a vanishingly small number.

BlackAdder wrote:Barry,
I presume you have read this page. http://wiki.laptop.org/go/OFW_FAQ
I read it to say that boot from USB should work if you have /boot/vmlinuz and /boot/initrd.img on the USB device. Might initrd.gz need to be changed perhaps?

HTH

I tried changing its name to 'initrd.imz' as that was in one of the examples, also tried uncompressed 'initrd.img'. No go.

Barry: Did you get a developer key? Also, I think to boot from the USB stick you have to turn the power on while holding down the "O" key on the game pad in order to initiate an "alternate boot".

Barry,

Maybe this will help:

Dual-boot Ubuntu/SD with Sugar - got it!
( http://olpcnews.com/forum/index.php?top ... 7#msg13017 )

delphi wrote:Barry,

Maybe this will help:

Dual-boot Ubuntu/SD with Sugar - got it!
( http://olpcnews.com/forum/index.php?top ... 7#msg13017 )

Ah ha! That might be it. I was getting that "Loading ramdisk" then it just froze. I'll try this fix.

BarryK wrote:I'll try this fix.

Barry,

Hope you can make Puppy work on the XO - that would be fantastic !

Best of luck - from a fellow Sandgroper

Shame that there was no feedback to questions from BK or raffy on BK's blog. Ominous that AMD has now disbanded and dispersed its Geode design team. VIA has to step into the breach to prevent another walkover for Intel?

Sage wrote: to prevent another walkover for Intel?

I think Intel will have quite a way to catchup - the Classmate is so far behind the XO's design [1] that it's not even funny - just see the coolling vents (just to mention the most obvious thing) on the Classmate and think about the conditions these laptops are supposed to be designed for...

As it is, the XO's hardware is 'good enough' to run most of the software comfortably - checkout the video showing concurrent multimedia usage using Slackware(?) with FWM desktop[2]...


[1] bunnies:studio, OLPC XO-1
( http://www.bunniestudios.com/blog/?p=218 )

[2] olpc-usb-boot
( http://tinypic.com/player.php?v=80m5z49&s=1 )

delphi wrote:
I think Intel will have quite a way to catchup - the Classmate is so far behind the XO's design [1] that it's not even funny - just see the coolling vents (just to mention the most obvious thing) on the Classmate and think about the conditions these laptops are supposed to be designed for...

I think its probably worth noting that Intel started shipping the Classmate in March 2007. Thats almost a year ago. Since then they've deployed 30 proof of concepts worldwide. So I think you can be sure Intel's progressed and learned.

Quote from the OLPC list:

We have painfully discovered the limitations of the mesh and current collaborative software in Mongolia, where the convolution of the number of laptops with bugs #5335 (more mDNS traffic than expected) and #5007 (mesh repeats multicast too much) make the perfect storm, which prevents anybody from using the network. We will continue to improve the mesh performance..

Batteries: .. the batteries are not lasting as long as expected. The extreme cold was the first suspect..

I think this is what happens with a "mass-produce-now-test-later" model. A conservative but more professional R&D model for OLPC should have helped it prepare for these post-launch "storms".

(Speaking of mesh, I work only with two teachers on cellphone "mesh" and I guess I am now short of time for it. Imagine a mesh of 500, such as in Mongolia. )

raffy wrote:Quote from the OLPC list:

We have painfully discovered the limitations of the mesh and current collaborative software in Mongolia, where the convolution of the number of laptops with bugs #5335 (more mDNS traffic than expected) and #5007 (mesh repeats multicast too much) make the perfect storm, which prevents anybody from using the network. We will continue to improve the mesh performance..

Batteries: .. the batteries are not lasting as long as expected. The extreme cold was the first suspect..

I think this is what happens with a "mass-produce-now-test-later" model. A conservative but more professional R&D model for OLPC should have helped it prepare for these post-launch "storms".

(Speaking of mesh, I work only with two teachers on cellphone "mesh" and I guess I am now short of time for it. Imagine a mesh of 500, such as in Mongolia. )

Ive tried explaining this problem of Mesh networking to many folks who jumped on the Mesh networks bandwagon. In literature its called a "broadcast storm problem".

My whole PhD was focused on finding ways to resolve this. Turns out everything works lovely in simulations, but not so in real life.

Cozybit folks who developed the OLPC mesh solution tell me they have been working on limiting the broadcast storm problem.

As with all engineering, nothing comes for free (unless you work at Google - food), the more you restrict or try to limit the broadcast storm, the less resilient your route discovery is... which in turn reduces the ability to find routes and results in performance degradation and yadda yadda yadda... (Ill shut up now).

However at some point there really is not much you can do without going multi radio and increasing the complexity. Maybe when we develop perfect radios.

iottmco wrote: Since then they've deployed 30 proof of concepts worldwide.

Sounds impressive . However, I'm not sure (and, unlike with OLPC, Intel's project details are not publicly available ) what 'concepts' they're trying to "proof" as most, if not all, of the technology in the Classmate PC is rather old - just repackaged...

iottmco wrote: So I think you can be sure Intel's progressed and learned.

I'm sure they will. After all, they have no choice - their whole 'Wintel' model they have relied on for their massive profits for so many years, if $200 OLPC XO-like PCs were to become common, will be in question. I expect they will scramble to produce a low-powered CPU to match the one in the XO first. However, to match XO's brilliant dual-mode screen is another story. And the list goes on...

raffy wrote:A conservative but more professional R&D model...

Wi-Fi technology is inherently complex one and there will always be problems in 'real life' depending on a particular scenario - I should know as I rely on wi-fi not just for my home network connections but for the access to the Internet as well. A "conservative" and "more professional R&D model" (not to mention billions of dollars) will not alway prepare you for all the possible problems. Just ask Intel - who can forget their, regardless of the controlled conditions available, spectacularly failed WiMax demo:

Intel red faced after WiMAX demo fails
( http://apcmag.com/5904/intel_red_faced_ ... demo_fails )

For those still waiting for Puppy on OLPC, in the meantime the following shows how easy it is to switch (ie without installing another Linux distribution) from default Sugar to a another desktop environment:

Bill Clementson's Blog - switching to Xfce on the XO
( http://bc.tech.coop/blog/080130.html )

MDD,

mdd wrote: I think the OLPC software really stinks.
...
I also had two spare eeePCs laying around, so I gave the two eldest the choice between the OLPC or the eeePC. The eee won hands down. It wasn't even close. --MDD

It easy to switch from Sugar XO interface to something more standard - see my post above. If your kids love, for example, reading (e)Books you may find that the XO is actually a much better machine...

delphi wrote:For those still waiting for Puppy on OLPC, in the meantime the following shows how easy it is to switch (ie without installing another Linux distribution) from default Sugar to a another desktop environment:

Bill Clementson's Blog - switching to Xfce on the XO
( http://bc.tech.coop/blog/080130.html )

Delphi

Thanks for the link to set up Xfce!
Ron

Ron,

rrolsbe wrote:Thanks for the link to set up Xfce!

You're welcome

There are more detailed instructions on the OLPC wiki ( http://wiki.laptop.org/go/Xfce ). If you (in the spirit of Puppy )would like something even simpler, and probably even faster, you could try (using the same instructions more or less) Fluxbox - just replace the yum line for xfce with the yum line for fluxbox.

Here are some screenshots of OLPC running under Fluxbox:

( http://www.flickr.com/photos/96666521@N ... 846937247/ )

More distros running on XO... :
Arch Linux and Crux Linux on OLPC XO-1
( http://olpcnews.com/forum/index.php?top ... 7#msg14387 )

Did Puppy on the OX hit a brick wall?

I am just wondering because I would like to reiterate how great puppy would be on this device, not only because it would be useful to people not using the mesh networking of sugar, but also for kids as they see that they need a more traditional environment out of the XO. I would love to put puppy as my primary OS on my xo, it would make it so much more useful to me!

Thanks for all your hard work, and hopefully something comes out of this

mbs348