Post a Comment
The biggest technical challenge for Microsoft is likely the Flash MTD (Memory Technology Device) that serves as persistent storage. It lacks the FTL (Firmware Translation Layer) that typical mass-market Flash devices employ to make them look like ordinary rotating block devices (i.e. hard disks) to the OS.
Therefore, the OS has to be aware of the Flash architecture, such as immutable erase blocks, in the I/O stack, and it must have a copy-on-write filesystem designed for sane metadata maintenance in light of these contraints.
Linux has a device driver for MTDs and a few filesystems specifically designed for Flash MTDs, including JFFS2 and LogFS. Although Windows CE runs on ultra-mobile Flash-based devices, I think that all of them use CF, SD, or some other kind of FTL-based media.
What Microsoft apparently begged OLPC to do was to include an SD slot on the motherboard so that integrators could ship the XO with an FTL-based Flash medium specifically to accommodate Windows. The onboard MTD would not be accessible from Windows.
Any list of key OLPC XO features has to include the MTD. The FTL is just a stopgap, and it's largely responsible for the lackluster write performance of today's SSDs. Ultimately, the MTD is the right way to implement solid-state storage, and it takes a project like OLPC to cut through the stagnancy of monoculture.
Microsoft should make their OS work with the hardware rather than make the hardware work with their OS, especially since their approach is clearly regressive. Computers shouldn't always have to be "Designed for Microsoft(R) Windows(R) XP(TM)". For once, at least, Microsoft should design their OS for whatever hardware is in demand.
I'd love to see Microsoft implement MTD support in Windows so that we can move toward solid-state storage in a sensible way. But for Microsoft to elbow their way onto the XO by insisting on legacy hardware accommodations is outrageous. How are we ever going to move forward given Microsoft's interest in holding us back?
Wouldn't that be nice. A return to OSes designed for specific hardware.
It boggles the mind how far we've strayed, to the point where hardware is now designed around software.
Oh, for the good old days. Back when software was re-written to work on older hardware, to eek every little once of performance out of the existing hardware.
Throwing more hardware at crap software is not the way to increase performance.
Really, in my opinion, this is defeating the object of the laptop. It has a customized Linux OS that has a very different user interface. It has been designed with certain goals in mind.
By putting XP on, which is *not* customizable to the same extent, you are removing one of the main parts of the project. The fact that you have to hack it (by placing an extra SD slot) should be your first warning.
Does Microsoft's greed know no limits? Why not just focus on getting XP on the Eee PC and leave this non-profit project alone?
"""
Nope.
All your base are belong to us!
"""
Indeed. And how many dollars per unit will buyers pay for that otherwise unnecessary SD slot? Multiply that by the number of units and what will it come to? Does the Windows tax know know bounds? Even users of hardware explicitly designed and built for another OS have to pay it!
I see it as just the opposite. The purpose of the OLPC was to have an open platform so that anyone could create an OS for it. By having XP on it will increase the amount of software available for the laptop tremendously. This may even increase sales of the laptop. As for the SD slot, thats one of the best features of the laptop according to those who have it, because they can load their applications onto the card and take it with them or order more cards and have unlimited storage. I see no greed here on Microsoft's part, just an interested party in working with the OLPC. After all, it is an open platform.
No, the fact that the specifications are available for all the hardware makes it open. Everything that's needed to write an OS from scratch for this device is there.
Basically the OLPC shows off the biggest problem there is out there. Microsoft's OS's successes are based on EXCLUSION: proprietary lockin. Now they're facing a market which is based on INCLUSION and they're failing miserably.
Now OLPC have to buildin new hardware just to have some crappy OS waste it.
It's like every shitty terminal around using Windows 95 while a mobile phone from yesterday would have enough power to do the job. No, you need a full blown x86 to have some dubious software running inside Internet Explorer. Be lucky if it runs through the whole year.
Probably the biggest criticism about OLPC is the "crappy OS". The UI is easy to use, but it is dumbed down and "toy"-ish. If one of the goals is to get kids into computers, it doesn't make much sense to provide a UI that works nothing like any other UI on the planet. On the flip side, it also makes ALOT of sense to provide the os that is used by 85% of the planet.
Don't get me wrong, there is alot to like about OLPC. Just not suger.
The point is not to train the next generation of Microsoft techies. The point is to get people access to information and teach them about computers in general, and how to program them to do what you want. Sugar is great for that. XP would be a huge waste.
What I am saying is more praising alternatives to sugar then saying xp is a must. IMHO sticking KDE on there would be best. Linux enourages experamentation, and if you use kde you wont be completely lost when you move to a "real" computer, be it linux, mac, or windows.
Edited 2007-12-06 17:10 UTC
While KDE is a lot more memory efficient than most people credit it for, I would not really want to run it plus several apps (including a web browser) on 256MB with no swap partition, nor sacrifice a large chunk of the puny 1GB flash disk
It's got some great Educational apps, though.
I think Sugar is on-balance about the best-suited GUI for the project. Also, I (and many people from my generation) grew up with computers that bear practically no resemblance to Windows and I don't think my generation has been at all disadvantaged by this: Windows is, after all, not particularly hard to use. And who is to say that the KDE or Windows from 10 or 15 years from now will bear any real resemblance to their present-day counterparts?
Don't underestimate the learning speed of children.
They will start out with the XO at the age of 7, later at high school use a normal PC with Linux or Windows on it.
Some of these kids will become software or hardware engineers, others will become farmers, musicians, medics or whatever.
But all of them will have seen at least two different interfaces! That alone is reason enough to NOT make the XO interface like KDE, Gnome or Windows.
I remember my start into computers: A sharp pocket computer which could be programmed in basic and had 16 kB memory and one line for display. I learned that, and had a start advantage when I later had to learn about PCs and programming. Re-learning gave me the chance to correct my partially wrong assumptions about computers. Re-learning is good for kids.
Compare looking at DaVinci's Michelangelo only from one side versus getting views from different positions. Which one is better?
You don't need to teach 7 to 16 year old children "Windows" or "KDE" or "OpenOffice". You need to teach them "Computers". And force them to look at the topic from different angles.
"The point is not to train the next generation of Microsoft techies. The point is to get people access to information and teach them about computers in general, and how to program them to do what you want. Sugar is great for that. XP would be a huge waste."
Yeah, because they won't be exposed to Windows at some point during their lives. And do you seriously think that people want to know how to program computers to do what they want? People would rather pay other folks to do that...not because they aren't capable of programming, but because they simply aren't interested.
And do you seriously think that people want to know how to program computers to do what they want? People would rather pay other folks to do that...not because they aren't capable of programming, but because they simply aren't interested.
Yes, you're a low class troll and yes I'm going to respond anyway. This laptop is not for the next generation of dipshit PHBs. It's for the next generation of folks who the dipshit PHBs will be giving all their money to.
I think MeatLoaf said it best when he sung, "you took the words right out of my mouth". I've seen a whole generation of people who have gone backwards in terms of IT know how. Anyone who has worked on an IT help desk can tell you that people, rather than learning how to use a computer have simply got 'macro's' programmed in their brain to get a task accomplished.
When things go wrong, when the software is changed, since they don't have the underlying fundamental understanding of what they're doing - they're confused, they can't do anything. When they move office suites, operating systems and so forth, rather than simply adapting, they have to learn everything again. It is the equivalent of changing ones shoes and having to relearn how to walk again - its stupid.
For those who will reply to me and exaggerate "oh, end users aren't meant to need to know all the technology behind it" and "oh, people aren't interested in programming" (god knows why jayson.knight has to ride on that hyperbole - learning programming has nothing to do with what people like me are trying to say) - that isn't what I am saying. What I am saying is that when they do something in a given piece of software - what is actually happening. It is about knowing the terms; what is an icon, what is the desktop, what is a menu - I've come across users who don't even know those basic concepts. I see them using computers and I wonder whether or not they're simply randomly clicking in locations on the screen in the hope of actually clicking on the thing they need.
What the world needs is for those people who do get the technology by way of OLPC to learn the technology the correct way. When their nation moves forward, the IT educated population is agile of the mind so that when changes to the IT infrastructure, there isn't major disruption and costly retraining.
Edited 2007-12-06 20:19
This is what really pisses me off about "computer courses" in high school and college. They don't teach people concepts (how to structer a document, how to structure a memo or a letter, how to do things). Instead, they teach you the combination of button clicks that will perform an operation (click here to bold, click here to center, click here to create bullets). It annoys the piss out of me how many people know how to use MS Word, but could writer a letter in WordPerfect to save their lives since the menu structure is different.
What is wrong with learning concepts that can be applied to any piece of software?
Why are people so lazy that they need to learn the magic button sequence for doing something, instead of learning the concept behind the "something" such that they can transfer that knowledge to any software application, OS, device, or whatnot?
Try getting a teacher to teach a course on document structure, and you'll get a bunch of teachers asking you where the MS Word step-by-step instructions are. It's shameful, to say the least.
Indeed. Who is to say that Microsoft will be behind the dominant user interface ten or twenty years from now, when the kids using these laptops will be productive adults? In time everything changes, and Microsoft's domination of the OS market will come to an end one day. Personally I'm hoping that either there is no monopoly after that day, or if there is, it is something like OSX that is based upon open software concepts. Even if the actual user interface is patented and closed, perhaps the hardware itself and the software backend will remain open and able to be changed by the users to suit their wants and needs.
Hello? It is designed for little children. It is way way better to have a open OS where kids can learn EVERYTHING about it than to have a nearly obsolete 7 year old crippled CLOSED OS.
If kids learn linux they wont have any problem using Windows ( they prolly dont want to but if they have to they can ). Children are smart they learn much faster than adults and they can adapt really fast. So there is absolutely no need for XP on OLPC. Everyone who thinks so prolly lost the ability to adapt.
Kids aren't going to learn linux, they are going to learn sugar, which does its best to keep linux hidden. As I said in a previous comment, kde would be ideal, but even xp will help more in actually developing useful skills. Just because it has linux under the hood doesn't make it pro.
Sugar has a feature built right into the interface that opens the source code for any application. That's got to be the most transparent user interface ever devised, trumping "view source" in the web browser.
How does Sugar "keep Linux hidden"? By not being KDE or GNOME? Come on, you'll have to do better than that on this site...
In any case, the OLPC objective is more about free software than Linux in particular. So if it emphasizes to these kids the importance of choosing free software in general rather than merely choosing a free software kernel, then they're doing the right thing.
Linux is often the most suitable kernel for any given free software product, but it's not enough to have the growing and youthful middle classes in developing nations worshiping Linux as the key to achieving independence from the multinational vendors.
The whole stack must be free, end-to-end, in order for these people, their businesses, and their governments to bring themselves into the digital cloud age while retaining their cultural identities and regaining their economic independence.
The kernel probably isn't the package that will see the most development from the XO community in the short-term. The work is likely to focus on Sugar, the core applications, and other user-visible components.
So it's important that Sugar is clean and hackable. Although KDE, especially KDE4, is pretty hackable for those who grok C++, it has more of a learning curve. It's not clear how useful it would be on a machine with such meager hardware and screen real estate.
This is some delusional Western imperialist mindset. Who gave anybody the right to go and define what it means to be a "pro" at doing stuff with computers? Give these people, especially the kids, some time with open hardware and open software, and they will surely develop amazing skills and functionality.
These people deserve better than figurative "Windows" into the conventional wisdom that locks them into the indentured servitude of global capitalism. They deserve an equal playing field that offer limitless possibilities for innovation and empowerment.
KDE and GNOME have a lot of our Windows-tainted conventional wisdom and cultural DNA already baked in. They already have their visions more or less in focus within their development communities. I think it's better for everyone to give fresh minds a fresh whack at a relatively clean slate.
They will show us that "pro" comes in all shapes, sizes, colors, and income levels. Let us have a little more faith in humanity, and let those of us living in relative comfort resolve for the coming new year to consider the economic, cultural, and spiritual value of bringing hope and opportunity to the lives of the world's vast underclasses.
Edited 2007-12-06 20:10
RE[3]: How ridiculous
Being forced to give up your work may be a choice a fully-informed adult can consent to and live with, but how will those countries who have bought in to OLPC react once they find out that the copyleft licenses Sugar is based on requires them to give up copies of everything their children have made to anyone who asks?
What?
I must have missed something there, but I wasn't aware of any free software licenses that extended to any content created with that software. I find it hard to believe I'll have to give away the .gif animations I've made because I created them in the Gimp, or the papers I wrote in OpenOffice.
The GPL, for instance, requires you to make modifications to the software available if asked, or the source code to programs derived from GPL'd code. I don't recall seeing any notice requiring all content created on an OLPC to be distributable under a Creative Commons license, even.
The OLPC is meant to be a learning tool. Yes, the operating system is open source, and modifications will also have to be open source... but Negroponte's vision for the OLPC was as a platform for a whole range of interactive educational and developmental* activities, not just computer programming. I'm more worried about how much of those curriculum activities and e-textbook materials actually exist, and I think we've been paying too much attention to what OS the hardware is running. Yeah, it's Linux, get over it.
*in the psychological and sociological sense
Edited 2007-12-06 22:11
Yet more ill informed nonsense bordering on trolling.
Quite aside from the other inaccuracies in this post that have already been addressed we have:
How are foreign kids going to benefit from being able to look at obfuscated, ill-documented C code with comments, variable names method names, class names, and library names written in English? And not even full English sentences, mind you, but a horribly contracted version (HorCntrVers) that's barely readable even to a native English speaker?
The OLPC's Sugar environment unlike the Linux kernel is not written in C but rather in Python. Python is an excellent language for learning programing. As for the use of English, you have to learn some English to program, as virtually all programming languages including those developed by non first language English speakers are written in English anyway. Learning English is educational too.
Hello? The kids aren't getting the laptop to learn about operating systems and how computers work, they are getting them to help educate them about reading and writing and arithmetic. It doesn't matter if the software was opened or closed for those tasks. The kids aren't learning programming for the age group this is targeted at. I don't see how if they learn linux they can learn Windows but not vice versa. Your arguments make no sense.
Could be that the GUI is not the best. I only saw videos of it. But at least they are trying to develop one which fits the needs best. How these needs are defined is debatable.
First, you don't need to switch OS to replicate the Windows look&feel.
Next, doing so makes no sense for this laptop. Wether or not the whole world uses Windows XP now, it won't use it in five years. And we all can really hope that some serious progress will happen in these years, because nowadays UI's arent what I would consider as very nice fitting to the needs of the user.
Having "learnt" Windows will be just useless then. If MS doesn't do radical changes in the next years, it will just loose. So in either case, today XP is just obsolete for an educational PC.
Part of the equation is OLPC's insistance on open source (software wise) and open specifications (hardware wise). The projects goal is - if I interpret and understand the mission statement correctly - not necessarily to train tomorrows computer users but to provide a machine, that helps to bring education, literacy, numeracy - short knowledge - to more children in challenging environments. Sure, the fact that it is a computer (and a very well thought out one) opens a range of possibilities (mesh networking, using it as a communication tool in rural areas with poor communication infrastructure, a light source, that is rechargeable by hand, etc.) but it's foremost aim is to enable more children to participate in the education process.
Education process >>> computer lab course.
The before mentioned openness of the OLPC guarantees, that once the need for a different User Interface emerges, it can be developed and deployed, even locally . The Asus Eee (quick, how many e's are there in this name - I always forget it) PC is arguably a less open platform (cf. initially undocumented modifications to the acpi / Wifi drivers for example) and people have nevertheless started almost immediately to scratch their itches, to install different distributions, etc.
My take is, that once the need for a more conventional user Interface emerges (e.g. either when the children that use the XO's grow older and start to learn more about programming, Office work, etc. or when countries start to buy XO's especially with the need for such more advanced courses in mind), there will almost certainly be a large number of projects that can and will bring IceWM, the Equinox Desktop, WMers like Flux/Black/Openbox, XFCE, stripped down versions of KDE or GNOME, Enlightenment, .... and probably a host of other, original and to-be-developed Desktop - Environments / Window managers to the XO.
EDIT: wording and fixed typo
Edited 2007-12-06 18:08
The goal is *not* to get kids into computers, it is to provide a general education tools to kids, you know, to help them learn about math, language skills, history, collaborative work, and so on.
Not only that, but UIs are easy to understand. You don't get "spoiled" because you've used one UI, i.e. it doesn't make other UIs harder to learn. Video games (who each have their own distinct UIs) have taught us that much.
I personally think it's great that this device does not railroad users into the Microsoft mold. Also, Sugar isn't that bad, and it's certainly not "crappy" - in fact, once you get used to it, it's kinda cool. In any case, you can install other Linux DEs on the OLPC if you're so inclined.
I like office a lot, but one word you cannot use with it is small... it uses a lot of memory, and a lot of disc space. Trying to get Office to run on something with the XO's space seems to me to be a fool's errand.
If they want to get an OS running on the OLPC's restricted space, why don't they use the embedded versions of Windows and Office designed for cell phones and such, as opposed to the full desktop OS? That way, you'd already have an OS designed to run in the low memory/power overheads of the XO, and with an office suite equally sized to match.
It's not like the XO is upgradeable or repairable; it's more of an embedded device anyways.
My first reaction as well. If resources are a problem then try to use XPe or CE instead of full grown XP. But if they're not going for such an obvious solution then the reason is clear. They want XP in there in order to open up OLPC to mainstream Windows desktop software, whereas Windows CE and XPe are niche platforms. In short, they see cheap, small, low power laptops as an emerging market and want to conquer it early.
I can understand why Microsoft are pushing XP, but surely it would be smarter to push MS Works than Office.
I know Works lacks a lot, but essentially is has most of the basic tools required for doing the job without a lot of the bloat. Surely the aim should be providing a complete, yet portable slimline package rather than the all or nothing attitude MS are pushing on this particular device?
Disclaimer: I'm neither associated with the OLPC nor with it's subsidiary in my country, OLPC Austria. But the place and topic of my work (university, particle physics) have been associated with the proverbial "ivory tower" a few times too often to let the following pass uncommented:
The OLPC XO is a real, existing (as in non-vapourware) and - judging from my limited experience with it ( half an hour with a prototype machine ) - very useable machine. Calling it "dead" and not useful is something that I certainly can't subscribe to.
The hardware alone is very innovative and seems to be *very* rigid. I had the pleasure to use a Panasonics Toughbook in the past and while the XO is certainly not as durable as the W2, for the money it costs it seems to be very reasonable. Combining the technologies that are used by the XO together in one *low price* machine and tell the world, that such a thing is not useful is in my book similar to people who highly over-estimate for example NASA's budget (it is less than 1% of the overal budget of the US, btw.) and complain about lofty, ivory-tower space exploration projects, that offers no real use (while ignoring a whole load of secondary by-products, that have proven to be incredible useful).
If you ask using fuzzy attributes ("decent" -> What do you mean with that?), you will get fuzyy answers.
One of my old boxes (256MByte Ram, Duron 600), runs Slackware with XFCE fine. Vector Linux, DSL and several flavours of popular distributions (Mepis antiks (?), Fluxbuntu, for example) provide alternative Versions of their products that aim at Hardware in the <= 256MByte Ram section.
If eye-candy is your thing, then I can recommend one of the e17 - beta based distributions. I had little problems with them on my ancient box.
(Then again, I ran Win95 on a 486er with 8Mbyte RAM so I'm tolerant).
EDIT:
As for space, I'm not sure if it would be a good idea to use OpenOffice in it's current shape on a machine with such specs. Something like Abiword / kword, gnumeric etc. is probably better suited for educational purposes and should -thanks to ODF support getting slowly better - allow for documents exchange with other computers, that run OO.
Edited 2007-12-06 19:14
Pepper Linux (embedded Linux distribution) can be run on XO machine. There is even an XO version although dated (version 3.1).
http://www.pepper.com/solutions/hardware_solutions.html
Speaking about office application, Write activity is based from Abiword.
Why in the heck isn't Microsoft modifying Windows Mobil to work with XO? There are already Mobile versions of Word, Excel, and other Microsoft products in addition to hundreds of freeware titles.
Even if they can get XP operating system trimmed down enough to install on the XO, what XP/Windows software can be used on it? Probably nothing...
Too bad Microsoft seems to feel the need to improve their products only when they're in a tight spot, in a market where they struggle. It says a lot about their complacency in the desktop market. When they're not challenged they crawl to a stop and just milk what they've got. I bet that if the likes of Linux and OpenOffice and Mozilla or Opera didn't come along there would be no Explorer 7 or Vista or Server 2003 or actually improved Office versions. In spite of any bullshit you may hear, it's still good old honest competition that drives innovation, not patents and monopolies.
RE[2]: creating slimmer version of XP is a Good Thing.
Hardly.
With the sudden arrival on the scene of the XO, the Everex gPC and the ASUS EeePC, all of them low-spec inexpensive machines running Linux, and all of them at least initially being apparently very popular purchases, Microsoft can see a market beginning to develop that they are not a part of.
The ASUS EeePC appraently expects to sell nearly 4 million units alone next year, and that is just one model.
Suddenly, a market for software for these devices will open up. It will be a software market that MS has no control or influence over.
That is what they are deathly afraid of.
You will note that they are busting a gut to make sure that there is an XP option for each of these machines.
The funny thing is, people will find they have to additionally spend a significant percentage of the cost to get XP running on these machines ... and they end up with a bare OS that can't do anything. Then they will have to spend an additional amount, possibly more than the whole original purchase, just to get the machine as functional as it was before they broke it by installing Windows on it.
It won't take long for the word to get out then. "Just leave the Linux stuff on the machine. It works way better, and costs nothing".
This is the lesson that Microsoft do not want people to learn.
Edited 2007-12-06 23:35
It won't take long for the word to get out then. "Just leave the Linux stuff on the machine. It works way better, and costs nothing".
This is the lesson that Microsoft do not want people to learn.
I forgot to mention ... if people put XP on the machine, and the start looking for safe secure functional free application software (in order to keep the costs from ballooning) ... they will suddenly find that the firefox + thunderbird + abiword/gnumeric or openoffice solution they arrive at ... is all available for Linux in the first place.
Then the penny will drop.
"We don't need the Windows, and if we keep the Linux and run the same application software anyway, we don't need anti-virus and we won't have to put up with WGA & DRM & activation keys and other similar nonsense either".
Ars Technica:
http://arstechnica.com/news.ars/post/20071205-microsoft-feeling-hea...
... has apparently arrived at more-or-less the same conclusion as I did as to why Microsoft is doing all this. I think like Ars Technica that it is fairly clear that Microsoft see a threat emerging.
http://arstechnica.com/journals/linux.ars/2007/07/19/intel-launches...
http://arstechnica.com/news.ars/post/20070710-hands-on-with-a-proto...
Edited 2007-12-06 23:58





