The One Laptop Per Child organization will use Linux on its inexpensive machines, but the operating system suffers the same code bloat as Windows, the project’s leader said Tuesday. My Take: A few months ago I blogged about this as if I knew what was coming. I still believe that the $100-laptop project should be targetted as an embedded application and so Qtopia with ARM is a better/cheaper/faster solution than Fedora/RHEL with x86. If Palm is able to sell the Zire 22 at $99 and still make lots of profit (yes, they do), then it is probably feasible to manufacture and market my suggestion at $100.
Negroponte: Slimmer Linux Needed for USD100 Laptop
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
Linux systems suffer from design bloat, Microsoft’s systems are suffering from something else.
Modularity isn’t free.
Of course, Linux itself doesn’t suffer from this TMK.
I experienced the Linux bloat problem when I wanted to set up an old machine to be a simple mail server. In the end I got FreeBSD 4.x to run on my 486-25 laptop with 12MB of RAM, 512MB HD and no CDROM drive. It runs fine as a simple mail server.
I have FreeBSD 4.11 on a P166 with 16MB of RAM and 4G of hard drive running a Squid proxy, freePops, SVN, postgresql and DNSMasq.
I think the $100 laptop probably has more power than those two machines combined.
Whoah, so now Palm resells Sharp’s products eh? You don’t say!
First off, a PDA is not a laptop. I am sure that Honda can sell a motorcycle for $2000.00, but that somehow does not mean that they can sell a car for that price.
Pretty badly researched piece of opinion IMHO, one should be more careful at researching the subject before giving any condescending opinion on what it’s or isn’t possible.
I guess Eugenia meant Zire instead of Zaurus.
Prices aside, I’d suspect that some type of embedded linux would actually work really well on these things – esp. in terms of battery life and startup/operational speed.
Hey, if you were actually READ my bloody blog post you would not reply to me this way. NOBODY said anything about PDAs, I specifically say that my suggestion in my blog is NOT a PDA, but a HANDTOP with a keyboard and a non-touchscreen LCD, like the Libretto was.
Did you actually read my blog I linked to, or you just read the 4-lines I wrote here at OSNews and you felt like replying anyway?
Edited 2006-04-05 01:41
You only need to look at how desktop Linux runs in its current state on 400MHz/128MB RAM class machine. GNOME might be able to pass as usable on such a machine, but when you add OpenOffice.org and Mozilla to the equation, you’ll start running into some serious performance problems.
This is really just a case of Linux being chosen because it has mindshare, and because the OLPC project has the financial backing of Red Hat. If you want software that isn’t half-assed in terms of the end-user, you’re going to run into problems here, because generally the Linux desktop (and by that I mean GNOME, Mozilla and OpenOffice.org) weren’t built with performance in mind. A lot of cuts to functionality and usability will most likely have to made for Linux to be a reality on the $100 laptop, and that isn’t good.
Personally I think they should have looked closely at all their options. If they demand open source, I think they’d be better looking into the likes of Haiku or perhaps even Syllable. Haiku may be incomplete, but in terms of being suitable for the $100 laptop, no more incomplete than Linux is and at 14MB, the whole OS is smaller than Linux’s kernel alone, which considering the $100 laptop is only going to have at most 1GB of storage, a very good thing. If they can accept a proprietary solution, as much as I hate to say it, they should really look in to either QNX or Windows CE.
And the cost. Oh yea, that wasn’t important on a $100 laptop. [/sarcasm]
You mention the problem with a Linux-based GNOME system is that Mozilla and OO.org are too heavy, but as alternatives you suggest operating systems that can’t even run them at all.
GNU/Linux should work fine on this laptop. It runs on less capable systems today.
Also, the article headline mentions “the operating system suffers the same code bloat as Windows”. Negroponte did not say that at all, at least that wasn’t quoted in the article.
Fedora Core is a fat distribution for the laptop because it’s not a system designed to work on “slim” (slow) computers. But because of its modular nature it can easily adapt and Red Hat is doing just that. I bet anything the OS will be ready way before the display, which is this laptop’s biggest problem right now.
You have a >14MB kernel image? How did you manage that?
kernel-2.6.16-1.2084_FC5.x86_64.rpm -> 15.6M
kernel+modules. Most of them could be stripped off though.
Because the hardware is fixed and resources are low, kernel for such a system may be monolithic to reduce the duplicated code.
I would like to have AROS laptop, but unfortunately it needs a HUGE amount of work to be usable and of course AROS authors should change priorities from backward compatibility towards modern features.
GNOME, Mozilla and OpenOffice.org
I hate to point out the obvious but I could have told you ages ago that the menioned 3some would not run comfortably on a 500MHz CPU with 128MB of RAM. Fortunately linux being as customizable as it is, there are other light weight alternatives to those 3 that will get the job done within those specs. Damn Small Linux as an example.
Haiku and Syllable suffer from incompleteness and a scarce software library that relies heavily on ports. Linux was a good choice, the kernel itself is well known, well documented, and can be easily tuned to scale down to small devices. This does not apply to Windows, and Windows CE has a pretty poor application library. Mindshare was not an issue.
The problem that exists is not with the operating system therefore, it is the application suite.
Personally I think KDE and KOffice would be a good bet here: KOffice is very lightweight, is more complete and integrated than “Gnome Office” (Abiword, Gnumeric and some other tools) and should be reasonably easy to pick up. Likewise, Konqueror is more lightweight than Mozilla, and is better integrated with the desktop. Lastly, several contributors in Africa have been working quite hard on creating translations for KDE in African languages (http://dot.kde.org/1004383235/ and http://translate.org.za/) including XHosa, Yoruba and Zulu.
The problem of course, is that KOffice is still a bit shaky. KDE 4, which is seeing a lot of rewrites across the board, should fix that, but it won’t be available till mid-2007.
>The problem of course, is that KOffice is still a bit
>shaky. KDE 4, which is seeing a lot of rewrites across
>the board, should fix that, but it won’t be available
>till mid-2007.
I don’t understand why people confuse rewrite with more stable, if anything a rewrite means a introduction of new bugs. And please don’t give me but qt and c++ means easily stable code, that is the mantra of KDE users for the past 4 years without the stability to show it.
KDE has the stability, feature-set, and resource usage to show that Qt/C++ was a good pick. Consider how easy it was for KDE to change their file-selector in comparison to the six months’s work it took Gnome to change theirs, or the problems Gnome had until last year with applications not respecting the global toolbar preferences entered by the user.
The problem with KOffice is that it has a tiny number of developers and a large and very old code-base.
Qt4 offers a lot more to developers, which reduces the work that individual developers need to do, making them more productive. Thus, when KOffice is re-written in Qt4, they will be able to ditch a lot of custom and old code for the well tested (tested by every other KDE developer!) features offered by Qt. While this could be characterised as a port, the depth and extent of the changes being made means that “rewrite” is a better term for what they have set out to do.
Lsstly, rewrite doesn’t necessarily imply more errors. While errors will occur, the prior knowledge generally serves to help developers avoid a lot of the semantic errors that had to be tweaks in the previous version. Further, that prior knowledge allows developers to write cleaner code, which again helps reduce the amount of semantic bugs.
I agree with Eugenia here… there is a lot that can be done if you look toward embedded applications. Also, linux distros like pdaxrom and openzaurus work great with small space.
They might even look toward bsd… I’m sure netbsd runs on whatever they have…
agree with the idea of Qtopia or GPE and the like, if they wanted an open source environment for it. if not, QNX would be the way to go.
Damn Small Linux fits on a 50MB disk, surely it wouldn’t be too heavy for this machine?
After installing Linux on my iPod Nano, I have a hard time believing they couldn’t put it on the OLPC machines.
DSL does not have an easy to use interface. It’s just a compilation of simple apps and front-ends. It does not have an underlying infrastructure and a unified API/toolkit like Qtopia, Gnome or KDE do.
Compression and disk space are not primary concerns. DSL runs extremely slow overall and eats up gobs of memory while providing mediocre software. There are much better examples for light Linux distros, btw, but…
Negroponte is absolutely right. Since the beginning it has been blatantly obvious that the only reason Red Hat Linux was selected for the project was because Red Hat was a major sponser, not because of any technical merits of the project.
I’d like to hope that this would be the kick in the pants that the GNU/Linux community needs to get its act in gear. But what we’ll see is more of the same: FUD.
Pot calling the kettle back, all I am trying to say that you are building a straw man argument regarding the pricing. Your blog entry does not change that fact… and yes I really “bloody” read it. Just because you list a bunch of parts does not mean that you can get the same price as a Palm. A product’s cost is not just a simple extrapolation of the compounding costs of its bill of materials… Just because you wrote a blog about it does not make you an authority in the matter by a long shot.
Plus this project aims for a $100 laptop, and you are proposing a handtop…. I guess whenever someone asks you for oranges you give them apples?
True, however I wasn’t suggesting to actually use DSL as the distro of choice, I was merely using it as an example of what you can do with very little memory. Qtopia would of course be a better choice in this context.
However, remember that any interface risks being alien to many of the kids who will use these laptops, as their computing experience will often be close to nil. That’s something to keep in mind.
Edited 2006-04-05 02:08
Bloat is relative. For example, people complain that 10 years ago, software took a lot less disk space and ran faster. Yet, they fail to mention that 10 years worth of features that *they* have asked for. Things that most people are starting to take for granted and would refuse to do without such as:
* International text input and display support
* Resolution independent text display (truetype fonts, anti-aliasing, etc.)
* Pixel perfect layout control
* Advanced grammar and spelling analysis tools
…and so on. Bloat is relative. One person’s bloat is another person’s must have feature. All of these features that many people would consider “basic”, didn’t really exist years ago. What’s funny is that the same group of people that complain about bloat in software, refuse to use competing software that has less features because their “must have” item (which others may consider bloat) isn’t implemented by that product!
You are right, but exactly those four elements do not create bloat.
Many OS’es have such functionality without being larger than OS/2 was in 1993.
The bloat typically stems from the “layer upon layer upon layer upon layer upon layer upom…” disease.
Ah, but those four features alone can substantially increase memory and CPU usage. I think many Linux users will remember the first time they used grep on a unicode-by-default Linux box when grepping across thousands of files compared to the speed of an older Linux distribution doing the same that used the POSIX LANG=C locale.
The poster gets modded to oblivion and you lash back feisty?
Most comments and activities related to this device have emphasized only single Linux OS types for use in the unit.
I propose using a number of Linux types, one built in OS of suitable size, and a number of LIVE CD linux types, such as Puppy Linux, DSL, Slimmed KNOPPIX, and others.
Many extra programs would be possible if such an arrangement were included, and not all repetitious of each other.
This assumes a CD is included……………
A possible setup at school for the students would be to provide several LIVE CD flash elements to be loaded by the instructor as desired with flash loading stations for the students laptops.
Several LIVE CD OS’es could be studied and utilized thereby.
A bare minimum of storage would be required to enable the special needs of the LIVE CD OS boots. (Actually now flash memory elements) since no CD is provided (too much power).
From the perspective of the embedded linux space that really made me pause to wonder what he could possibly be looking at. Probably misquoted, that or maybe just laying the groundwork for the price increase and switch to Windows Vista. “Linux was too fat, Bill showed me how to make Windows thin”
I deeply appreciate Nicholas Negroponte’s efforts to make $100 laptop for students. This is not a project to make profits but to give to poor students may be at a loss.
Students are not a inferior lot. Every generation they are smarter than their parents and it has to be that way for human evolution.
Never ever think its because they are students their requirements are not as sophisticated as adults. Students’ requirements in most cases far exceed adults’ requirements.
Its good Negroponte decided to use a Linux on these laptops. I personally use a flavour of Linux named Tomahawk Desktop (http://www.tomahawkcomputers.com/). This is a new multimedia Linux. I call it poor man’s Apple.
Lets see what I do with it. I use Tomahawk Desktop to browse Internet, send and receive email, write documents using OpenOffice, create graphic using Inkscape, create PDF brochures, copy artworks to thumb drive and give to printer to make copies, connect my digital camera and transfer photos to the computer, and use Gimp to touch up photos and burn to a CD and give it to a photo shop to print, I listen to digital radios, podcasts, I put a audio CD in and convert songs to mp3 and transfer to my mp3 player.
Isn’t what students need to do using their computers whether they are poor or not? Isn’t that what they should expose to whether they are poor or not?
Computer is not a calculator, that time passed very long time ago, now you can use it do lot of useful things and you can use it to learn lot of useful things.
If lot of students in a classroom do have many laptops, they can further do projects, isn’t it? They can create their school website? They can create artworks for competitions, etc. May be they earn money for their school by creating posters, artworks, logos, websites for others specially for businesses, etc.
As a side benefit of the laptop project, once students leave the school, they come out with lot of skills in their hand. They can easily find a job or can start a small venture without being a trouble for their parents or for their governments.
When you give something, whether it is free or at a fee, you should give something useful. You may still give these students a laptop at $100, but it should really be useful, not just operational.
I’d like to hope that this would be the kick in the pants that the GNU/Linux community needs to get its act in gear. But what we’ll see is more of the same: FUD.
What does Fear, Uncertainty and Doubt have to do with it?
Please note that FUD does not equal “propaganda”, it is a sub-category of it.
The other FUD. ****ed Up Disinformation.
I disagree that linux — the kernel — is bloated. I put together a distro that’s about 1MB, fits on a floppy disk, and gives me dumb-terminal emulation on a 486 laptop that I have (20 MB Ram). I also have a 2.4 kernel on my Sharp Zaurus 5500.
Now, once you put xwindows + (gnome or KDE) with all the necessary libs on top of a linux kernel, you’re basically not going to run it well on a machine with 128MB ram. I agree that somethine like qtopia or some sort of non-xwindows frame-buffer gui toolkit would be good.
If haiku or syllable were further along then either one would be a viable alternative.
Anyway, the bloat is not in the linux kernel. I wouldn’t even say that gnome or kde are especially bloated; they’re just not designed with embedded systems in mind. Calling them bloated is not fair to the developers (who, e.g. gnome 2.14, are making strides in improving performance).
I also agree that, while DSL is a fine hobbyist/hacker distro, it’s not a unified, newbie-friendly environment.
I’d personally like to see something like xfce (new gtk improvements should make it faster) + some fb xserver. I also think openoffice 2 should not be considered — it’s enormous. It would be better to go with some simpler tools with fewer features.
Windows CE!
(Just kidding…but it did kick ass on my HP Jornada with 16MB RAM in 1999. The device had no hard drive, only cost a couple hundred dollars, had a 56kbps modem, VGA out, a PCMCIA card slot, USB, ran Pocket versions of Word, Excel, IE and Outlook, and is still extremely useful today. In fact I just sold it to a college kid who wanted to use it to take notes in class. I’m sure HP could make the same thing with the same specs for $100 or less today.)
Gnome is not the best solution to develop on. Sure its cheaper for the developer, but its not as intuitive as the other popular desktop, KDE. You can avoid alot of bloat because KDE’s base would make OOo unnecessary. There is a complete office suite available in the KDE project. I’ve run IceWM on a paltry P120 non-MMX machine with only 20MB of EDO RAM and it was fine for basic tasks. Perhaps they can take the IceWM shell and add the KDE office features to it. Its not like this $100 has to sing, dance, and perform 1001 party favors.
Edited 2006-04-05 03:38
I’d actually agree about the icewm part. I’m running it here on my old 350Mhz G3 as it provided the best environment while being zippy enough for a comparitively low spec’ed machine. My seven year old son doesn’t have a problem with navigating it either which I’d think should put to rest the notion that it wouldn’t be ‘user friendly’ enough or whatever. Kids adapt, it’s adults that get more set in their ways.
As to the office suite part however, I’d think something like Abiword et al would be more appropriate.
http://www.kaisersite.de/dfm/
DFM adds desktop icons to IceWM, making it even more friendly to the average joe end-user. This would be a simple way to get a icon driven desktop with the small footprint of IceWM. Its simple, easy to configure, snappy as any GUI ever, and doesn’t hog memory. All good reasons to use it. A $100 machine shouldn’t be a Windows XP replacement.
I agree.IceWM is an excellent window manager.But how about Nautulis has the file manager?I tried that combo on my Pentium 3 600 MHZ computer (which is close to the $100 laptop’s specs,by the way),and it is fairly speedy.
If desktop linux would have had a coherent, intelligent strategy then it wouldn’t have had to be a bloated mess.
Who’s that -Linux guy- you keep talking about?
G.
This guy is complaining that a full regular linux environment is too fat, and then claiming that WinCE is much lighter. Of course WinCE is much lighter, as it is made for embedded environments. Linux too can be reworked down and put into a solid, embedded environment. Granted, an off the shelf distro will not work, as to really get the most out of your hardware, you need to highly customize the build process of what you are using, and thus your system needs to be built by hand.
And another thing, WinCE is not Windows.
“Suddenly it’s like a very fat person (who) uses most of the energy to move the fat. And Linux is no exception. Linux has gotten fat, too.”
No, it’s not fat. You chose to wrong distro for your project. You could have chosen any number of distros aimed at embedded and low resource environments but instead you went with RH/Fedora and GNOME. Not that there’s anything wrong with either of those but neither is suitable for the target environment.
It’s not Linux fault that you made the wrong decisions.
For sure, hard to believe, can he be that big of an idiot? Linux is whatever you make it in the embedded space, you roll your own distro. He rolls a fatty and blames either the kernel or generally slams every distro big and small, (what does he mean by, “Linux”?)he’s either mis-quoted or a complete ass, pick one.
Last time out I put Xfree, ice and netscape 4.72 in a 32mb cramfs.
I’m running RH9 on a P1/200MMX/64MB machine (~1GB drive) and works just fine.
Granted, it’s not as fast as my dual – dual core Opteron machine, but it works just fine as multimedia station.
RedHat can decide to detach the OLPC from the main Fedroa tree and rebased on older, tighter version.
Again, I see no reason why RedHat can’t pull it off.
Oh… and considering the fact that the last Vista test release I tested (a couple of weeks ago) needed ~600Mb to display the desktop, I find the “Linux is bloated” propaganda hilarious.
G.
What should have been done was ether using a small distro that’s already made (puppie linux, DSL, or even Knoppix. Or (what I would have done) installed LFS and created my very own disto just for that system then cloned it and copied it to all the $100 machines. It sounds like someone was just too lazy to take that extra step.
I think some readers press the panic button too early. Update status about OLPC can be found on http://fedoraproject.org/wiki/OLPC and https://www.redhat.com/archives/olpc-software/
Speaking about that quote:
Not that there’s anything wrong with either of those but neither is suitable for the target environment.
Watch the demo simulated with QEMU. http://people.redhat.com/berrange/olpc/sdk/olpc-qemu-admin-demo.htm…
A project called RULE Project http://www.rule-project.org demonstrated the ability to install Fedora on old PC. Last year, I managed to run Fedora Core 4 on a Pentium 133 with only 64MB using WindowsMaker and Fluxbox as desktop environment. Of course, heavy applications like OpenOffice have to be stripped out. Amazingly, the system runs decently until the old system died due to hardware failure.
Nokia with its Maemo DE is participating in the project FYI, http://wiki.laptop.org/wiki/OLPC_software_task_list
If OLPC didn’t meet the deadline, then you can panick.
I guess we all already know Linux GUI land is a total bloat. I hope this criticism is taken positively.
ok comparing qutopia to a desktop OS is just irresponsible. I am not sure that you actually have seen what the Zaurus can do and as a person that owns one I can tell you that it is nowhere close to being a PC replacement. The same goes for any pda that I have seen. They are a very fancy organizer that can also play games and do fancy stuff like view pdf file word and other popular formats ( with limited capabilities ) and has limited editing capabilities. While a 400Mhz might be kinda slow for a modern computer it still doesn’t mean that you can’t run a free OS on it. I think they would have to reconsider the part about having only 128mb ram plus ram is actually cheap when you are not looking for high performace stuff. As far as linux being bloated goes … well yes it probably is but so is any other software that exists today as most programers care about features first and performance 2nd. Hey there is not a single college in the world that teaches CS majors anything but that. And why shouldn’t they? Computers preddy much double their performance every 6 – 12 months. The problem with the $100 PC project is that first of all those guys had no clue as to how they are going to achive their goal. They had only though. As a result it’s been 2 years since they started and yet no product to show. I expect them to take at least another 2 years before the product is finished. Only at that point the hardware specs and pretty much every aspect of the platform would be ridiculously old. These guys are going agains every business principle and the whole thing is starting to look a lot like those organizations that dump old computers to developing nations and call it a donation.
This is like comparing the VW Polo with the BMW 7 series. The polo is light but slow compared to the 7 series, which is heavy but fast. How come? The engine in the BMW is way stronger than the one in the VW.
You can see the CPU as the engine of the laptop, and the OS as the body-parts. A car with a strong engine (fast computer) can carry a heavy body (Fedora Core 5, WinXP, Mac OSX) and still outperform the lighter car ($100 laptop w/ Lightweigth linux distro).
What they try to do is put a VW Polo engine in a BMW 7 series body. This car will perform like shit. Likewise will a laptop with a slow CPU perform with FC5 on it.
So i suggest they stick with the basics and make a custom linux distro that offers only the things really required. Like wifi, abiword, custom educational software and xfce or similar WM. There are a lot of things (OO.org, FF 1.5, Beagle, Mono, etc.) in FC5 that are nice for people with fast pc’s, but are not required by people in Africa for instance.
But you can go with a Polo to exactly the same places as with the BMW, and carry the same people.
Eugenia’ s shown specs sound like kids using it at school yet. Not for learning, but for gaming, music and video. At the moment a GP2X is ready in 10 seconds and it can run Qtopia too. Never read s.th. on OSNEWS. The gp2x community builds up breakoutboxes to use thinks like usb hdd, wifi, mice, keyboards …
specs like
TV-Out (S Video)
Audio Out (L+R)
RS232 Serial Connection Port
3 x USB Port (Host Function)
EXT Port Connector
DC 12V In
Power LED
JTAG as standard.
A break through cable to power up GP2X. No need additional PSU for GP2X.
A dipswitch to select USB powered by GP2X or PSU.
http://wiki.gp2x.org/wiki/Main_Page
http://www.gp32x.com/board/index.php?showtopic=27289&st=0
http://www.artaylor.co.uk/~richard/gp2x/
I guess it’s a question of being pragmatic. The OLPC project has to work with the sponsors it has and the material it has, not what an ideal world would choose.
A little more worrying is the “top down” nature of the project, at least as publicly presented. I wonder how thoroughly they’ve gone round their target audience in Africa, South-East Asia, South America, etc., and tried to establish what the folks there want. Delivering the wrong stuff probably means it just won’t get used.
All big projects involve challenges that have to be really sweated. In this case, coming up with a slim and fast OS is clearly one. I guess Negroponte is going to have to fight hard to get it. I still think proper Linux is the right choice, because it will bring folks up on a fully viable OS used all over the world. This will give users computer skills for later in life, a career and qualification path if they want, etc. Qtopia sounds too limited to me unless “embedded” is what careful research shows the target audience wants.
It’s a pity that Microsoft and Intel are now bandwagonning this idea with concepts of their own and spreading criticism and discord. It’s pretty sick watching the Wintel corporations manoeurving around philanthropy when profit centre and lock-in is what they really mean. If they wanted to help, they could start by asking what they can do for OLPC and not what OLPC can do for them.
Given the limited space and the limited write cycles on flash devices, I guess there will be no swap space?
Marry that to the 128mb of RAM, and that is going to be really difficult get going smoothly!
Edited 2006-04-05 10:08
I have not seen any offical statement about the CPU being a x86.
Remenber that AMD also produces MIPs clones which would fit the target power consumption (2 watts for all the system) much better than the geode x86 processor.
According to the OLPC wiki, it is a Geode x86 processor.
I think Zeta, BeOS and Haiku should be run well on this computer…
Yes. I also thought of Be (in whatever flavour). Or maybe QNX (HK might agree to use it for very fair (read: free for non commercial usage) conditions).
If a ppc was chosen I’d suggest something like MorphOS. Sure, it has its constraints, but it has a pretty small footprint, a *full* installation (with a very comfortable and powerfull graphical user interface) is about 30 MB. Even smaller than Be. It is the fastest OS out there I assume.
With the Freescale MPC5200B SoC there would be a good starting point for a cheap low energy reliable design IMHO.
Fedora is one of .. if not the heaviest out there …
Minimum RAM for text-mode: 128MiB
Minimum RAM for graphical: 192MiB
Recommended for graphical: 256MiB
Search for “embedded” in distrowatch & there is a loooong list
Vector is a great Linux BTW ;P
they need to re-evaluate some of their core decisions. Obviuosly GNOME is inappropriate, but their are other desktops that can accomplish this well. I am rather partial to XFCE as it is a full desktop environmnet (and not just a window manager). It can run on systems with as little as 32 MB of RAM according to an XFCE based distro called Luit Linux. OOo is to big (I don’t recall anyone in the project ever recommending it) so you need something like Abiword (which is becoming more appealing as they work on ODF support). They need to look at the hard numbers comparing FF to Epiphany to whatever else seems appropriate. DON’T JUST BUNDLE FIREFOX BECAUSE IT’S POPULAR! After that nothing else is as important because their are GOOD XFCE applications for most other things (mousepad, xfmedia, and xffm which will soon be replaced with thunar). The other things that will help these students are terminal applications like gcc, ftp, and ssh. The one largish item I would add is Evince for PDFs.
All that should be able to fit on OLPC. Those that claim this isn’t user friendly enough I would like to immediately disagree with. These children will quickly learn these tools, which is part of the goal:
Learning to use computers, not learning to use the lates WIN32 clone.
It took me all of a few minutes to come up with this, so I am confident that RH and OLPC will come up with an acceptable solution.
I’ve searched on the inet, but I can’t seem to find a list of stuff they plan to put in it.
I see people bleating about Gnome and Firefox and OpenOffice, but that simply can’t be true given the specs.
http://osnews.com/permalink.php?news_id=14231&comment_id=111617
have lots of links for you, even a demo.
And no, there are no OOo nor Firefox.
I even wonder why people talk about these apps, when people said it would be Gnome, which already has a browser and some office products.
http://osnews.com/permalink.php?news_id=14231&comment_id=111617
have lots of links for you, even a demo.
And no, there are no OOo nor Firefox.
I even wonder why people talk about these apps, when people said it would be Gnome, which already has a browser and some office products.
Actually, the demo shows Firefox as being installed and not Epiphany. It also doesn’t show any office software at all. There is a general expectation of some sort of proper word processor other than gedit or vim will come with it. This is why everyone is discussing the issue.
http://osnews.com/permalink.php?news_id=14231&comment_id=111617
have lots of links for you, even a demo.
Yes, lots of link, but nothing usefull to me.
as we know there are many window manager other than gnome and kde
they can provide same functions that gnome or kde can do, but in more high speed and use more less resource
such as Enlightenment DR17
its total size < 40M and is high perfermance
so I think if he use linux kernel + xorg + other window manager program,
he can meet he needs
What made it more mobile was the crank.
Now, without a crank and without a tailor-made OS, i fear the gadget will be obsolete (even for USD100) before it can see the light of the day.
What made it more mobile was the crank.
The crank has been replaced with a foot-pedal, which is much more ergonomic.
Whatever robust and ergonomic the fact remains a foot-pedal is a separate device, while a crank is an integrated device, that makes a huge difference in mobility. The crank idea was the best innovation and an integral part of the hype in the richmen ecologist world. I agree it doesnt change much from a practical use point of view, nevertheless the best gift to a poor child is something that is valued by the richman.
Remember Commodore trying to invade eastern europe with low-cost C64s ?
The morale is poor people don’t want all-inferior products. The product can be cheap with low specs but it must have its own special something, otherwise obviously it worths nothing, it would be dismissed as a gift from “white” people.
After all, Bill Gates can be trusted when it comes to what consumer want, so i fear the MIT philanthropy will miss the african market.
Edited 2006-04-06 00:03
I’m running Linux on an HP iPAQ with 64M RAM and 32M Flash. It’s got a 200MHz ARM CPU. It runs the QT-based Qtopia environment just fine, and the GTK+ based GPE just as well. It lets me listen to music, watch videos and pictures, manage my agenda, write notes, program in Python, read ebooks and play games.
Nokia’s tablet is just a bit more powerful and has a larger screen, and it surfs the internet just fine. It also runs AbiWord and Gnumeric, which are pretty full featured office apps.
With twice the RAM, 16x the storage and 5x the processing power, i’m sure the OLPC will be able do many more things, and do them well. No, it won’t run the full Fedora Core 6, but neither does it have to. This is an EDUCATIONAL device, meant to deliver schoolbooks and teach reading and math skills to children; OpenOffice.org is neither necessary nor desirable here.
The idea to use Windows CE is laughable: why choose a proprietary, expensive OS with few apps? Not to save hardware resources, that’s for sure: my iPAQ does far more things with Linux than it ever did with Pocket PC.
All other alternatives, like QNX, Syllable, Plan 9, VxWorks or OS/X are either too weird, too expensive or both, while offering almost nothing valuable in return. Mindshare is important; Linux will afford the project the free work of thousands of programmers worldwide.
My take is that the final distribution for OLPC will look a lot like Nokia’s Maemo; it is lean and sexy, and a very easy target for the Gnome and GTK+ apps that Redhat favors. And above the basic “Desktop” functionality, a wide range of educational software will have to be developed, which is something that misteriously nobody is speaking about.
Right. I suggest you find your own distinctive acronym and try to popularize it, instead of muddying the debate.
My, my, it’s like listening in to a bunch of high-school chicks in the changing room —
does this distro make me look bloated?
It’s not my fault, I got this desktop as a legacy from my parents!
This program isn’t so fat, it just looks that way until you turn it sideways . . .
I think those who believe that Red Hat is going to pay $$ to get their distro on this device, and then try to stuff a regular desktop package into essentially high-end handheld hardware are mistaken; I think arguments over the relative bloat of Linux vs. Windows or POS vs. Qtopia vs. Wince are way off topic; I think the issue of what OS/distro would best fit the OLPC hardware specs in an ideal world is interesting and worth discussing but largely irrelevant to the actual situation here.
I think Negroponte’s comments were a public part of the negotiations between him and Red Hat about how much time and resources will be spent modifying the Red Hat components to suit the project. Instead, I’m thinking how neat it would be to have a distro that’s ultra-lite, lean, and runs fast on this type of platform, and I hope RH’s thinking this way, too.
SC