Fedora Core, a popular flavor of Linux being developed by the open-source community, is hardly hefty by today’s operating system standards. But to fit the tight specs for the $100 laptop being designed by MIT’s One-Lap-Per-Child group, Fedora will need to go on a crash diet, concedes its overseer, Red Hat. The Linux software maker confirmed today that it has donated $2 million [EUR 1.65 million] to the OLPC and joined as a corporate member and said it plans to put some of its brightest engineers to the task of slimming down Fedora before the laptops debut in early 2007.
Wouldn’t it be simpler to build a custom Fedora version? This almost seems like a waste, since we already have proven technology for that kind of thing. Nonetheless, I believe it will be worth it.
Uh…correct me if I’m wrong, but didn’t it say that they are building a custom Fedora version? Right in the article summary in fact.
Perhaps you meant to question why they’re going with Fedora as opposed to something else?
Uh…correct me if I’m wrong, but didn’t it say that they are building a custom Fedora version? Right in the article summary in fact.
No, the article says that they’re gonna “slim down Fedora” and Fedora “will need to go on a crash diet”, not that they’re building a custom version.
Maybe create a kind of Fedora Lite using Kadischi(a live CD generation still on early stage) or use light version of some applications. Some services will need to be disabled for faster bootload.
As for ease to use frontend package manager, I recently tested Pirut (replacement for system-config-packages) on Rawhide. I can say that PM is simple, intuitive. One of features will allow user to add his/her own repository.
Starting from FC5, the release schedule is back to every six months.
I like this, mabe they can pull a firefox!
So the poor kids will get a computer with an OS that is non maintained after 6 months….poor kids!
If it does all they need, and is reasonably well-integrated and bug-free, what will they care?
From the looks of the specs mentioned in the article and other things I’ve read this isn’t going to be a normal Linux operating system anyway. Unusual and slower hardware (the screen, from what I’ve read, is particularly weird), smaller memory and storage space, requirements for language localization and special capabilities… This thing is probably going to be more like an appliance.
I doubt there will be massive hardware differences to require updates for, at least unless/until they decide to make the OLPC v2. Unless something in the software goes massively wrong, there’s really nothing to update.
If it does all they need, and is reasonably well-integrated and bug-free, what will they care?
One team is working on mesh networking for the MIT-$100 laptop, and providing mesh functionality is one of their goals. I guess that if there is a hole in the kernel, or some commonly used software with networking functionality, they’ll want to patch it.
Quite a nice thought: (signed) updates that are distributed through mesh networks.
So the poor kids will get a computer with an OS that is non maintained after 6 months….poor kids!
That’s why special arrangements are probably made.
All in all, it is a very good direction. Red Hat has been out of the marketing buzz spotlight lately, but they are pushing some very interesting projects:
– Systemtap, a framework easing collection of information about the running kernel for perfomance and functionality diagnosis.
– Frysk, a system monitoring and debugging tool. The use cases show some neat examples: http://sourceware.org/frysk/usecases/
– A free (as in freedom) Java platform, based on GCJ plus classpath. Works very well for non-Swing, and non-AWT applications. A good match with SWT or gnome-java.
– Speeding up YUM, and they wrote a up2date replacement for YUM, pup.
– The Fedora directory server.
– The GFS cluster filesystem.
All in all very interesting for developers and enterprises. Over time Fedora/RHEL, and other distributions, have grown quite heavy as a result. Not much of a problem for enterprises, but it is for kids having $100 laptops, or other people on a tight budget. It is good to see that they are addressing this.
Hopefully they will also take in account (if possible) that kids in a third world country can have huge potential if the tools that we use are also available to them. That means that they have to take it further than just providing a stripped-down version of Fedora.
My hopes are quite high
.
Red Hat has been out of the marketing buzz spotlight lately, but they are pushing some very interesting projects:
Not to mention stateless Linux. The could be the killer approach to getting Linux on the corporate desktop.
Well if we cut all the development stuff (gzz, headers etc= an drop the server sw (and maybe even drop kde and gnome and uuse something light (xfce etc)) I guess we can trin down fc quite a bit, and they can still ude the normal fc updates (sent to the schools on CD/DVD in areas wiyhout broadband. Is this a good starting point?
PLZ comment.
Why not just modify damn small linux? It can be configured to be very flash memory friendly (MIT laptop uses flash not a hard drive), and it’s only 50mb. So, sure you might want to add openoffice, educational software, etc. But the core would be so small.
That sounds good, also why can’t they use the embedded linux being used by Nokia in the Nokia 770 Internet Tablet. It’s light and designed for a device of that size, power useage etc.
http://www.maemo.org/index.html
I think Red Hat got in there because of the money they invested.
Edited 2006-02-03 01:37
I think they will certainly take lot of bits from maemo project, because it is free software
I doubt that they will dublicate the efforts. Certainly they will work together on certain elements.
And, ohh, you can’t simply take it and cram it down to MIT Lap because hardware and architecture will be quite different. So they will take a bits here, and there, and propably more communication about embeded hardware will accour between devs. So I watch it with high hopes.
am i missing sumthing or might have debian been a better solution?? its pretty easy to slim up enuf to fit in the tinyest of spaces… with fully guing systems..