OK, so that's a tidy handful of smart and knowledgeable people. Thanks for answering the call. :)
I wanted to talk about my recent experiences in education-land. This is a fairly lengthy email, so apologies in advance -- but I think that the people here are the perfect folks to read it.
The question, for those who want to skip to the end:
What should Red Hat be doing in the education space?
===
Last month, I went to visit Mike Huffman in Indianapolis. He's in charge of technology in schools for the state of Indiana. He said he was going to take me out to show me how Linux worked in his schools.
I'd seen a bunch of cool Linux labs elsewhere. Good example: I visited Jeff Elkner at Yorktown High School (Arlington, VA) a few years ago and toured his K12LTSP lab. It was essentially a lab for teaching computer skills, in the wealthiest high school in a wealthy school district, with a highly motivated teacher. Impressive stuff, but not what I would call a broadly replicable success.
So on my visit with Mike, he took me to a school in Greensburg, Indiana. This was not a wealthy school; in fact, it was in an economically depressed area. A lot of parents lost their jobs in the past few years when the local plant shut down. Fairly typical story nowadays, it seems.
As we drove out to Greensburg, Mike told me the story of how he came to believe that open source in general was *the* solution to the "computers in education" problems in his state. He told me about how Microsoft was squeezing him at every turn, and yet how the computers he had were sorely underutilized. He really explained to me, for the first time, the ideas around one-to-one computing -- and why open source is ideally positioned to make one-to-one a reality in his state.
===
I was expecting Mike to take me to a computer lab. Instead, he took me to an English class.
The kids filtered in, chitchatting like kids do. When the bell rang, the teacher directed their attention to the URL she'd written on the board. The kids turned on the monitors mounted underneath their plexiglass-covered desks, fired up their web browsers, and got to work.
The URL was a Moodle quiz. Something about "The Red Badge of Courage" or something, I don't remember. (As so often in my school days, I wasn't paying attention to that bit.) But the kids were done with the quiz in, oh, five minutes. When they were all done, the teacher started to teach her class. The kids would occasionally Google something. The teacher had a supernatural instinct about which kids were working on class-related stuff and which kids were fooling around, and kept the class pretty well in line.
I talked to her after class. "Moodle and Criterion have saved my life," she said. "I used to spend hours grading papers and quizzes. Now, Moodle takes care of the quizzes, and Criterion grades the papers for spelling and grammar so I can focus on the content. This software saves me 10 hours a week -- which I spend building the actual curriculum."
When I asked her about how she created the content, she said "oh, I get help from the other English teachers; we build the lesson plans together." Whereupon Mike Huffman broke in and told me that this was one of the first lessons he'd learned: the absolute necessity of collaboration. When Mike put *one* lab into a school, that lab failed. The teacher was intimidated by the technology, wouldn't ask for help, and the computers would sit unused. But when he put *three* labs into a school, the labs prospered; the teachers compared notes, learned from each other, and ultimately took fierce ownership of these fantastic new tools they'd been given.
The next day, I went to the symposium for the teachers in the state of Indiana, and heard similarly breathless stories. I heard from a teacher of *twenty-five years* who said that her one-to-one lab changed her mind about taking early retirement. "I can focus on actual teaching now," she said.
The common wisdom that old teachers can't adopt technology is clearly wrong. If you give smart teachers the tools to do their jobs, they will use those tools. In fact, the veteran teachers will be *more* effective than the younger teachers, because they've got the classroom management skills to make it work. I've seen the proof.
===
All of this tells me that a lot of folks have been selling the whole "computers in schools" concept completely wrong. In Indiana, they are not, not, *not teaching computers*. They are teaching *kids*, and they are *using* computers to do it. It seems like an arbitrary distinction, but it is in fact a *fundamental* distinction -- and it's a distinction that so many people seem to miss. Until very recently, myself included. Sometimes you have to see these things firsthand to understand the impact.
So why don't teachers embrace technology? The common "wisdom" goes something like this:
"How can you expect a teacher to learn all this computer stuff when they've got all this other work to do, like grading papers?"
When the success stories go more like this:
"How can you expect a teacher *not* to learn all this computer stuff so they stop wasting their time on grunt work, like grading papers?"
===
So now that I've been converted, I ask myself: "what role should Red Hat play here?"
This is the hard question. Lately, it's the question that has been keeping me up nights.
I know that Red Hat has one hell of a commitment to education, because I've seen it. I've seen it in the late-night IRC logs of the OLPC developers who work crazy hours on a project that, two years ago, was widely regarded to be complete crackrock.
But OLPC isn't enough. For one thing, it's designed to solve problems that kids in the developed world don't have, and thus may be regarded by many in the developed world as an expensive toy -- especially when a desktop system can be had for $300 or less. (This is true even in some of OLPC's target markets; you wouldn't believe the resentment I heard from some Brazilians at being considered "too poor for real computers".) For another, it may not work in the classroom nearly as well as we think; if even one or two kids leave their laptops at home, suddenly you don't have a one-to-one classroom anymore. Most of all, though, it's just unwise to put all of your eggs in one basket -- and that is precisely what Red Hat has been doing in the worldwide K12 space with OLPC.
Meanwhile, there's this once-in-a-lifetime opportunity unfolding before our very eyes. Microsoft is leaving the door wide open by delivering a product that people don't want to buy. One-to-one computing is picking up steam. Open source applications like Moodle are changing the classroom workflow -- really, changing what is *possible* in the classroom.
Where is Red Hat? And where *should* we be?
===
We are very fortunate here at Red Hat. We have the best support brand in the open source business -- in fact, one of the best brands in the whole software business. The old saying used to go, "nobody ever got fired for buying IBM." In the new world of open source, it's becoming increasingly true that "nobody ever got fired for buying Red Hat."
The Shadowman logo opens doors for us -- and for open source -- in whatever market we choose to enter.
That's why it is so deeply frustrating to me that, for years now, Red Hat has watched from the sidelines as the K12 open source community has labored. Not to say we haven't made some efforts, because we have. We've hosted a mailing list here, an RHN channel there. We put together an open lab for North Carolina schools, with mixed success. One time, we printed up some *awesome* T-shirts. We helped Eric Harrison package up K12LTSP way back in the day -- what has it been, Eric, six years now? -- but instead of capitalizing on our relationship with him and taking the opportunity to build something more meaningful, we sat back and watched. And watched. And watched.
And still we watch.
===
So let me throw this question out to you all.
What *should* Red Hat be doing in the educational space? Something that makes us *just enough* money to justify a business case, but helps Linux advocates make a *real* difference in schools -- the kind that Mike Huffman is making in Indiana? Because I'm tired of us doing nothing, and I'm looking for any good advice that I can get.
Thanks for reading this far.
--g
On 4/16/07, Greg Dekoenigsberg gdk@redhat.com wrote:
snip........
What should Red Hat be doing in the education space?
I challenge you to call HP, Dell, IBM or Gateway and ask them for a quote on a K12LTSP lab for a school. Place the call directly to their educational sales rep.
People have a hard time using what they don't have and can't get.
Honestly, give it a try. Someone try it and report back to this list what happens.
K12LTSP motto: It works. It's free. Duh...
And the hardest part of that to explain is...
;-) Paul Nelson
On Mon, 16 Apr 2007, Paul Nelson wrote:
I challenge you to call HP, Dell, IBM or Gateway and ask them for a quote on a K12LTSP lab for a school. Place the call directly to their educational sales rep.
People have a hard time using what they don't have and can't get.
Honestly, give it a try. Someone try it and report back to this list what happens.
K12LTSP motto: It works. It's free. Duh...
And the hardest part of that to explain is...
So I've been talking with lots of folks both inside and outside the company about this. And what I hear from the inside folks is really interesting.
We've got a sales team that is, in fact, responsible for selling into schools. They're called the "FED/SLED" team -- FEDeral gov't, State/Local gov't, EDucation.
They've got a sales number to hit. I can't go into details, but suffice to say... it's not terribly ambitious. :)
And why is that? Two reasons, it seems like:
(a) We don't have that much to sell that *clearly* targets the space. (b) Even if we did, we don't have the manpower to sell it.
And the fact that (b) is a problem seems to prevent us from making serious headway on (a).
So now I'm starting to believe that the way forward looks something like this:
1. Find a way of associating the Red Hat name with worthy education projects. Of which K12LTSP is clearly one, but there are others.
2. Build a channel. Red Hat will never be able to service schools directly, ever -- and neither can Dell/HP/whomever, because even if they *did* sell systems with K12LTSP built in, they wouldn't realistically be able to service schools either.
But we *can* provide expertise to local partners who can.
Maybe we can provide strong mechanisms for giving channel partners the ability to go sell the crap out of "a proven solution" -- which is training for school admins, plus some low cost/high value level 2 support, plus some other stuff that is valuable for the partners. And in my internal discussions with Redhatters, this is now looking like a sensible goal to shoot for: channel enablement.
But wait -- K12LTSP works, and it's free! True enough. But Red Hat has lived in this paradox for a long time:
A lot of smart people, for some reason, just don't *trust* free.
In the big commercial markets, that's fantastic for us, and allows us to grow a great business and subsidize further development of Linux and open source.
But in the education market, this weird distrust of free-ness gets in the way of changing the way computers work in schools.
My goal now: to come up with a business model that makes Red Hat *just* enough money -- and it doesn't have to be much in the larger scheme of things -- to push forward the one-to-one message.
I'll be at NECC. Steve, I'll see you there. I'm sure I'll be talking first hand with a lot of you as well.
--g
Here's my 2 cents...
Just like many people use a Windows operating system, or the OS X operating system, or Fedora, Red Hat, Ubuntu, Suse or whatever, I don't think the OS is really that big of a deal. Nearly everyone who has ever sat down in my K12LTSP lab can use it after just a little clicking around. So I don't think Red Hat's strategy should be to promote their OS.
Speaking as a teacher, what I think is missing for other teachers is familiar resources they can use to do what they do. In Greg's original post, did those teachers just sit down and learn Moodle on their own? I kind of doubt it. We say that we should be teaching word processing and spread sheets, not Microsoft Office, good luck finding such resources. Yes they are out there, but not in the places most educators normally look, such as the website of the big textbook publishers.
With all this in mind, I think Red Hat, if they want to put company effort into making their product specifically, and open source generally, into the minds of educators is to promote things like: -publishing and selling textbooks that specifically teach word processing skills, not just Open Office or Word or Abi or whatever -same for spread sheets -same for photo editing -same for desktop publishing -web design -programming -Have available, for extra cost or free, Moodle course material for any of these courses -Provide low cost, highly available training in Moodle (or any other open source classroom software deemed appropriate). This could even be done as a train the trainer type sessions where a teacher who attended the training could then go back and teach others. -help develop moodle material for other courses of popular textbooks. As a math teacher, I've attempted to do Algebra lessons for Moodle, but with equations and graphs, you have to create images for all that. I imagine the same issues would exist for physics or chemistry.
I think the best approach would be indirect. Not promote the OS, but promote activities that would require or make desirable, open source.
That's my opinion. But as I like to say, I'm just a math teacher from Iowa.
Eric
On 4/17/07, Greg Dekoenigsberg gdk@redhat.com wrote:
On Mon, 16 Apr 2007, Paul Nelson wrote:
I challenge you to call HP, Dell, IBM or Gateway and ask them for a quote on a K12LTSP lab for a school. Place the call directly to their educational sales rep.
People have a hard time using what they don't have and can't get.
Honestly, give it a try. Someone try it and report back to this list what happens.
K12LTSP motto: It works. It's free. Duh...
And the hardest part of that to explain is...
So I've been talking with lots of folks both inside and outside the company about this. And what I hear from the inside folks is really interesting.
We've got a sales team that is, in fact, responsible for selling into schools. They're called the "FED/SLED" team -- FEDeral gov't, State/Local gov't, EDucation.
They've got a sales number to hit. I can't go into details, but suffice to say... it's not terribly ambitious. :)
And why is that? Two reasons, it seems like:
(a) We don't have that much to sell that *clearly* targets the space. (b) Even if we did, we don't have the manpower to sell it.
And the fact that (b) is a problem seems to prevent us from making serious headway on (a).
So now I'm starting to believe that the way forward looks something like this:
- Find a way of associating the Red Hat name with worthy education
projects. Of which K12LTSP is clearly one, but there are others.
- Build a channel. Red Hat will never be able to service schools
directly, ever -- and neither can Dell/HP/whomever, because even if they *did* sell systems with K12LTSP built in, they wouldn't realistically be able to service schools either.
But we *can* provide expertise to local partners who can.
Maybe we can provide strong mechanisms for giving channel partners the ability to go sell the crap out of "a proven solution" -- which is training for school admins, plus some low cost/high value level 2 support, plus some other stuff that is valuable for the partners. And in my internal discussions with Redhatters, this is now looking like a sensible goal to shoot for: channel enablement.
But wait -- K12LTSP works, and it's free! True enough. But Red Hat has lived in this paradox for a long time:
A lot of smart people, for some reason, just don't *trust* free.
In the big commercial markets, that's fantastic for us, and allows us to grow a great business and subsidize further development of Linux and open source.
But in the education market, this weird distrust of free-ness gets in the way of changing the way computers work in schools.
My goal now: to come up with a business model that makes Red Hat *just* enough money -- and it doesn't have to be much in the larger scheme of things -- to push forward the one-to-one message.
I'll be at NECC. Steve, I'll see you there. I'm sure I'll be talking first hand with a lot of you as well.
--g
-- Greg DeKoenigsberg Community Development Manager Red Hat, Inc. :: 1-919-754-4255 "To whomsoever much hath been given... ...from him much shall be asked"
Fedora-education-list mailing list Fedora-education-list@redhat.com http://www.redhat.com/mailman/listinfo/fedora-education-list
On Tue, 17 Apr 2007, Eric Brown wrote:
Speaking as a teacher, what I think is missing for other teachers is familiar resources they can use to do what they do. In Greg's original post, did those teachers just sit down and learn Moodle on their own? I kind of doubt it.
As I understand it, they went to workshops that were sponsored by the Indiana DOE. But again: in schools where a single classroom was equipped, that class did not use the computers. But in schools where multiple classrooms were equipped, the computers were wildly successful.
Teaching effective use of Moodle (or maybe Sakai) is clearly one of the requirements for a successful one-to-one computing initiative.
We say that we should be teaching word processing and spread sheets, not Microsoft Office, good luck finding such resources. Yes they are out there, but not in the places most educators normally look, such as the website of the big textbook publishers. With all this in mind, I think Red Hat, if they want to put company effort into making their product specifically, and open source generally, into the minds of educators is to promote things like:
-publishing and selling textbooks that specifically teach word processing skills, not just Open Office or Word or Abi or whatever
Let me ask this question:
Why textbooks? Textbooks are expensive, relative to bits. If the goal is to teach computer skills, why not make a computer-based curriculum the primary method of teaching?
- Have available, for extra cost or free, Moodle course material
for any of these courses
I'd say the Moodle curriculum should be the basis -- not the texts.
-Provide low cost, highly available training in Moodle (or any other open source classroom software deemed appropriate).
I agree with this. Some combination of CBT-based training and local knowledge.
This could even be done as a train the trainer type sessions where a teacher who attended the training could then go back and teach others.
Which appears to be exactly how Indiana is doing it.
-help develop moodle material for other courses of popular textbooks. As a math teacher, I've attempted to do Algebra lessons for Moodle, but with equations and graphs, you have to create images for all that. I imagine the same issues would exist for physics or chemistry.
Right. This is a clear issue: I'm seeing a divide between what's possible for English / History / Social Studies and what's possible for Math / Physics / Chemistry. Moodle seems to be more useful for the former disciplines. Looks like NASA is working on some interesting stuff in the latter, though.
I think the best approach would be indirect. Not promote the OS, but promote activities that would require or make desirable, open source.
Yeah, I agree here as well. The nice thing is that the open source alternative can be the "reference implementation" that just so happens to be free.
That's my opinion. But as I like to say, I'm just a math teacher from Iowa.
Math teachers seem to be underrepresented, so I'm glad to see your participation.
If you had your druthers, what software would exist to teach math?
--g
On 4/17/07, Greg Dekoenigsberg gdk@redhat.com wrote:
Let me ask this question:
Why textbooks? Textbooks are expensive, relative to bits. If the goal is to teach computer skills, why not make a computer-based curriculum the primary method of teaching?
I can do with or without textbooks. I just know some teachers feel like they couldn't exist without one. Even though I would rather work from a computer, there are people who could read books all day, but not a computer screen. I currently teach all my computer classes without a physical text book. This includes Computer basics, computer repair, Web design (html), advanced web design (PHP/MySQL), networking, and video production. Often, I point them to web resources, or free and open textbooks on the server.
I would personally embrace a computer based curriculum. The only downside would be when the server or network is down. Last week my server died, but I had a new one within 3 days. Luckly I had a computer related video to show that took three days.
If you had your druthers, what software would exist to teach math?
Personally, I would love to see an open source version of Scientific Notebook from MacKichan software (http://www.mackichan.com/). Its pages can be exported as HTML with images. My favorite feature is that it includes the MAPLE engine, so you can put in an equation, click a button, and it solves the equation. It does simple arithmatic to advanced calculus, 3d graphing, differential equations, the works. A teacher can also build quizes or test questions where instead of having one exact answer, a teacher can describe a range of random numbers to be used, and a formula to find the answer. For example, say a student is supposed to be able to find the slope given two points. the teacher may define the values for the points to be random whole numbers between -10 and 10, and define the answer using the slope formula, but the kid sees plain numbers. I've hunted and hunted for an open source alternative, but the closest thing I can find requires learning a new programming language, which I don't have the time for.
Eric
On Tuesday 17 April 2007 10:49, Greg Dekoenigsberg wrote:
On Tue, 17 Apr 2007, Eric Brown wrote:
Speaking as a teacher, what I think is missing for other teachers is familiar resources they can use to do what they do. In Greg's original post, did those teachers just sit down and learn Moodle on their own? I kind of doubt it.
As I understand it, they went to workshops that were sponsored by the Indiana DOE. But again: in schools where a single classroom was equipped, that class did not use the computers. But in schools where multiple classrooms were equipped, the computers were wildly successful.
You understand it correctly. If you like I can probably prod Huffman into joining the list.
Teaching effective use of Moodle (or maybe Sakai) is clearly one of the requirements for a successful one-to-one computing initiative.
We say that we should be teaching word processing and spread sheets, not Microsoft Office, good luck finding such resources. Yes they are out there, but not in the places most educators normally look, such as the website of the big textbook publishers. With all this in mind, I think Red Hat, if they want to put company effort into making their product specifically, and open source generally, into the minds of educators is to promote things like:
-publishing and selling textbooks that specifically teach word processing skills, not just Open Office or Word or Abi or whatever
Let me ask this question:
Why textbooks? Textbooks are expensive, relative to bits. If the goal is to teach computer skills, why not make a computer-based curriculum the primary method of teaching?
That would be wonderful and fabulous. The bottom line is there needs to be content. The biggest complaint I've heard about moodle so far is, "I have to type it in?"
- Have available, for extra cost or free, Moodle course material
for any of these courses
I'd say the Moodle curriculum should be the basis -- not the texts.
-Provide low cost, highly available training in Moodle (or any other open source classroom software deemed appropriate).
I agree with this. Some combination of CBT-based training and local knowledge.
This could even be done as a train the trainer type sessions where a teacher who attended the training could then go back and teach others.
Which appears to be exactly how Indiana is doing it.
Again, correct. But getting that initial person to buy in is a pain. Especially when other forces are driving them away from it for their own reasons/agenda.
-help develop moodle material for other courses of popular textbooks. As a math teacher, I've attempted to do Algebra lessons for Moodle, but with equations and graphs, you have to create images for all that. I imagine the same issues would exist for physics or chemistry.
Right. This is a clear issue: I'm seeing a divide between what's possible for English / History / Social Studies and what's possible for Math / Physics / Chemistry. Moodle seems to be more useful for the former disciplines. Looks like NASA is working on some interesting stuff in the latter, though.
I think the best approach would be indirect. Not promote the OS, but promote activities that would require or make desirable, open source.
Yeah, I agree here as well. The nice thing is that the open source alternative can be the "reference implementation" that just so happens to be free.
That's my opinion. But as I like to say, I'm just a math teacher from Iowa.
Math teachers seem to be underrepresented, so I'm glad to see your participation.
If you had your druthers, what software would exist to teach math?
--g
What should Red Hat be doing in the education space?
First off, I don't think that this should be looked at as just "education space". Many of the approaches to help education could also be sold to businesses. Businesses need to save money as well, and the education space leads to the business space after high school or college. So when putting together a model, it may help to look at the education space as a foot in the door to the business space.
Here is a quick background on myself. I work full time as the Tech Admin for a Senior High School, Junior High School and 3 elementary schools with about 850 students total. I also run a local computer business on the side geared mostly for commercial support along with some residential and 4 other local schools. I work with Windows, OSX and Linux.
I think thin clients and central system management is the way to go for schools and many businesses. However I see 2 major setbacks in this area. First is scalability, currently there is not a foolproof, efficient single setup that can scale many servers with hundreds of clients. I know that Jim McQuillan has talked to an organization in South America (I believe it is somewhere down there) who has put together an excellent load balanced multi server setup that can handle a few thousand client and he hopes to build this into future releases. But that isn't guaranteed and is definitely a way out in the future. There needs to be a fast and easy way to centrally manage 5000 thin clients with a single server cluster. This is the main thing keeping Thin clients out of schools in our area. It just doesn't make sense to manage a separate system in each classroom, they all need to be tied together so profiles and data are available in every classroom. Sure Samba/LDAP with NFS mounted /home directories are a way to get there, but this isn't always speed friendly and is not widely tested for hardware guidelines. Second is multimedia. Much of what students need computing for very multimedia based. Terminals (and fat clients for that matter) need to be able to deliver good and consistent sound for all apps. They also need to be equipped with the necessary software to access all types of online content. Many sites students use contain Windows Media, Shockwave, Flash, Real Media, Quicktime movies, etc. Schools need a distribution that does this out of the box and is stable doing so. They also need this all to play from within a web browser.
I am not an advocate of one to one computing being achieved with laptops. I have overseen laptop programs in the past, and do not think this is the way to go. They are costly to purchase, costly to maintain, and have a shorter life cycle than thin clients. I personally think that the best way to give students access from home is to have them remotely connect back to the server cluster at the school and work just as they did in school. This has been done for years in businesses and there should be no reason they can't do it in schools. However remote connection clients for Linux are lacking in comparison to their counter parts. Windows Remote Desktop can pass sound, communicate with local printers, and is much better at speed. If I use VNC compared to RDP I simply do not get near the same experience, RDP makes me feel as if I am onsite where VNC has menu and typing lags. I think an excellent Remote software package would be a great add-on as well.
So I guess to summarize I would love to see the following come about in the future:
1. Seamless Cluster Management. With Windows 2003 server I can fire up a management console and choose server roles, Linux needs this. I would love that when I fire up my new Redhat Server it asked me what roles I would like to assign to this machine. Roles could be Primary or Secondary server, Application server (gui to walk through which apps to export to other servers), DHCP server, File Server, Directory Server, DNS Server, etc. Then when I choose each option it walks me through a little wizard to configure the server role. If I happen to add a server and choose Secondary, it would allow me to pick which primary server to choose for each role with a simple gui that let me enter in the IP address or name of which server was the primary for File serving, Authentication, Applications and so on. I really think a nice GUI server role management tool would go a long way.
2. Strong Multimedia delivery. I know there has been a lot of talk that a move PulseAudio as opposed to ESD or ARTS should help. But along with sound multimedia browser integration and a single do all player would be great. I know multimedia is a must for schools out of the box.
3. A good compression remote desktop tool that can transport sound and recognize local printers.
4. RedHat is obviously closely partnered with Dell since RedHat is the only distro Dell will install on their servers. Work with Dell to package server setups that can handle different sized schools. If I could call Dell and talk to a server rep and let them know that I am looking at deploying a Redhat based thin client setup in a school with 800 terminals and the rep could tell me that I need servers XYZ and ship them as a configured package that I can just plug into my network and start adding terminals, I would be in heaven. This wouldn't go just for schools either, businesses could benefit from this as well. Right now when I mention such a thing to a Dell server Rep they haven't a clue what I am even talking about. There could be a few case studies and some testing to determine how much hardware is needed and what the most efficient way to cluster would be. (Personally I find a speed hit when using Samba/LDAP with NFS mounted /home directories, there has to be a more efficient way of doing things.) Redhat is definitely know as a stable server operating system, couple that with Dell hardware support and preconfigured packages, and you could stumble into a goldmine. Maybe if Redhat and Dell worked together you could find 64bit server setups with quad quad core Zeon processors and 32GB of RAM and 4 teamed Gigabit NICS that could handle 1000 clients from a single machine. People like me just don't have the resources to test this sort of thing and see if it is possible.
5. Work with Dell or some other vendors and build an approved client list that is known to work with the above preconfigured packages out of the box. This could provide a start to finish package that is guaranteed to work. Maybe Dell could even look at getting into the thin client market along with RedHat.
6. Be sure that all common Education apps are in RedHat software repository and can easily be installed with Yum Extender. I don't think that they need to be prepackaged, but need to be easily searched as Education applications and be available. A good scheduling/grading appliation such as http://richtech.ca/openadmin/ or http://www.miller-group.net/index.php?option=com_frontpage&Itemid=1 are a must in this list.
7. Possibly build local vendor support for the sale and technical support of such systems. I know my business would be interested in supporting Southern Minnesota, Wisconsin, and Iowa area. http://www.1-cs.com. Local support options may be a key in implementation and making end users feel comfortable. If nothing else knowing support is available could set minds at ease.
I think if Redhat could find solutions to the above problems they could dominate one to one computing and the terminal services market in both Education and Business. I am not sure how much of the above is even in the realm of RedHat's scope but if nothing else they may be able to be a good facilitator.
And you thought your message was long :-) Jim Kronebusch
On Tue, 17 Apr 2007, Jim Kronebusch wrote:
What should Red Hat be doing in the education space?
First off, I don't think that this should be looked at as just "education space". Many of the approaches to help education could also be sold to businesses. Businesses need to save money as well, and the education space leads to the business space after high school or college. So when putting together a model, it may help to look at the education space as a foot in the door to the business space.
Here is a quick background on myself. I work full time as the Tech Admin for a Senior High School, Junior High School and 3 elementary schools with about 850 students total. I also run a local computer business on the side geared mostly for commercial support along with some residential and 4 other local schools. I work with Windows, OSX and Linux.
I think thin clients and central system management is the way to go for schools and many businesses. However I see 2 major setbacks in this area. First is scalability, currently there is not a foolproof, efficient single setup that can scale many servers with hundreds of clients. I know that Jim McQuillan has talked to an organization in South America (I believe it is somewhere down there) who has put together an excellent load balanced multi server setup that can handle a few thousand client and he hopes to build this into future releases. But that isn't guaranteed and is definitely a way out in the future. There needs to be a fast and easy way to centrally manage 5000 thin clients with a single server cluster. This is the main thing keeping Thin clients out of schools in our area. It just doesn't make sense to manage a separate system in each classroom, they all need to be tied together so profiles and data are available in every classroom. Sure Samba/LDAP with NFS mounted /home directories are a way to get there, but this isn't always speed friendly and is not widely tested for hardware guidelines. Second is multimedia. Much of what students need computing for very multimedia based. Terminals (and fat clients for that matter) need to be able to deliver good and consistent sound for all apps. They also need to be equipped with the necessary software to access all types of online content. Many sites students use contain Windows Media, Shockwave, Flash, Real Media, Quicktime movies, etc. Schools need a distribution that does this out of the box and is stable doing so. They also need this all to play from within a web browser.
I am not an advocate of one to one computing being achieved with laptops. I have overseen laptop programs in the past, and do not think this is the way to go. They are costly to purchase, costly to maintain, and have a shorter life cycle than thin clients.
I've heard this complaint with laptops elsewhere.
I personally think that the best way to give students access from home is to have them remotely connect back to the server cluster at the school and work just as they did in school. This has been done for years in businesses and there should be no reason they can't do it in schools.
Indiana is also doing this. An interesting side-effect it has on students: when they log in from home, they don't conceive of themselves as "doing homework". They think of it as "finishing work at school." Which is a brilliant sleight of hand to accomplish in a kid's mind. :)
However remote connection clients for Linux are lacking in comparison to their counter parts. Windows Remote Desktop can pass sound, communicate with local printers, and is much better at speed. If I use VNC compared to RDP I simply do not get near the same experience, RDP makes me feel as if I am onsite where VNC has menu and typing lags. I think an excellent Remote software package would be a great add-on as well.
/me nods.
So I guess to summarize I would love to see the following come about in the future:
- Seamless Cluster Management. With Windows 2003 server I can fire up
a management console and choose server roles, Linux needs this. I would love that when I fire up my new Redhat Server it asked me what roles I would like to assign to this machine. Roles could be Primary or Secondary server, Application server (gui to walk through which apps to export to other servers), DHCP server, File Server, Directory Server, DNS Server, etc. Then when I choose each option it walks me through a little wizard to configure the server role. If I happen to add a server and choose Secondary, it would allow me to pick which primary server to choose for each role with a simple gui that let me enter in the IP address or name of which server was the primary for File serving, Authentication, Applications and so on. I really think a nice GUI server role management tool would go a long way.
- Strong Multimedia delivery. I know there has been a lot of talk that
a move PulseAudio as opposed to ESD or ARTS should help. But along with sound multimedia browser integration and a single do all player would be great. I know multimedia is a must for schools out of the box.
- A good compression remote desktop tool that can transport sound and
recognize local printers.
- RedHat is obviously closely partnered with Dell since RedHat is the
only distro Dell will install on their servers. Work with Dell to package server setups that can handle different sized schools. If I could call Dell and talk to a server rep and let them know that I am looking at deploying a Redhat based thin client setup in a school with 800 terminals and the rep could tell me that I need servers XYZ and ship them as a configured package that I can just plug into my network and start adding terminals, I would be in heaven. This wouldn't go just for schools either, businesses could benefit from this as well. Right now when I mention such a thing to a Dell server Rep they haven't a clue what I am even talking about. There could be a few case studies and some testing to determine how much hardware is needed and what the most efficient way to cluster would be. (Personally I find a speed hit when using Samba/LDAP with NFS mounted /home directories, there has to be a more efficient way of doing things.) Redhat is definitely know as a stable server operating system, couple that with Dell hardware support and preconfigured packages, and you could stumble into a goldmine. Maybe if Redhat and Dell worked together you could find 64bit server setups with quad quad core Zeon processors and 32GB of RAM and 4 teamed Gigabit NICS that could handle 1000 clients from a single machine. People like me just don't have the resources to test this sort of thing and see if it is possible.
- Work with Dell or some other vendors and build an approved client
list that is known to work with the above preconfigured packages out of the box. This could provide a start to finish package that is guaranteed to work. Maybe Dell could even look at getting into the thin client market along with RedHat.
- Be sure that all common Education apps are in RedHat software
repository and can easily be installed with Yum Extender. I don't think that they need to be prepackaged, but need to be easily searched as Education applications and be available. A good scheduling/grading appliation such as http://richtech.ca/openadmin/ or http://www.miller-group.net/index.php?option=com_frontpage&Itemid=1 are a must in this list.
- Possibly build local vendor support for the sale and technical
support of such systems. I know my business would be interested in supporting Southern Minnesota, Wisconsin, and Iowa area. http://www.1-cs.com. Local support options may be a key in implementation and making end users feel comfortable. If nothing else knowing support is available could set minds at ease.
I think if Redhat could find solutions to the above problems they could dominate one to one computing and the terminal services market in both Education and Business. I am not sure how much of the above is even in the realm of RedHat's scope but if nothing else they may be able to be a good facilitator.
And you thought your message was long :-) Jim Kronebusch
Heh. :)
Thanks for the feedback, Jim. I appreciate it.
--g
On 4/17/07, Greg Dekoenigsberg gdk@redhat.com wrote:
On Tue, 17 Apr 2007, Jim Kronebusch wrote:
What should Red Hat be doing in the education space?
First off, I don't think that this should be looked at as just "education space". Many of the approaches to help education could also be sold to businesses. Businesses need to save money as well, and the education space leads to the business space after high school or college. So when putting together a model, it may help to look at the education space as a foot in the door to the business space.
Here is a quick background on myself. I work full time as the Tech Admin for a Senior High School, Junior High School and 3 elementary schools with about 850 students total. I also run a local computer business on the side geared mostly for commercial support along with some residential and 4 other local schools. I work with Windows, OSX and Linux.
I think thin clients and central system management is the way to go for schools and many businesses. However I see 2 major setbacks in this area. First is scalability, currently there is not a foolproof, efficient single setup that can scale many servers with hundreds of clients. I know that Jim McQuillan has talked to an organization in South America (I believe it is somewhere down there) who has put together an excellent load balanced multi server setup that can handle a few thousand client and he hopes to build this into future releases. But that isn't guaranteed and is definitely a way out in the future. There needs to be a fast and easy way to centrally manage 5000 thin clients with a single server cluster. This is the main thing keeping Thin clients out of schools in our area. It just doesn't make sense to manage a separate system in each classroom, they all need to be tied together so profiles and data are available in every classroom. Sure Samba/LDAP with NFS mounted /home directories are a way to get there, but this isn't always speed friendly and is not widely tested for hardware guidelines. Second is multimedia. Much of what students need computing for very multimedia based. Terminals (and fat clients for that matter) need to be able to deliver good and consistent sound for all apps. They also need to be equipped with the necessary software to access all types of online content. Many sites students use contain Windows Media, Shockwave, Flash, Real Media, Quicktime movies, etc. Schools need a distribution that does this out of the box and is stable doing so. They also need this all to play from within a web browser.
I am not an advocate of one to one computing being achieved with laptops. I have overseen laptop programs in the past, and do not think this is the way to go. They are costly to purchase, costly to maintain, and have a shorter life cycle than thin clients.
I've heard this complaint with laptops elsewhere.
I personally think that the best way to give students access from home is to have them remotely connect back to the server cluster at the school and work just as they did in school. This has been done for years in businesses and there should be no reason they can't do it in schools.
Indiana is also doing this. An interesting side-effect it has on students: when they log in from home, they don't conceive of themselves as "doing homework". They think of it as "finishing work at school." Which is a brilliant sleight of hand to accomplish in a kid's mind. :)
However remote connection clients for Linux are lacking in comparison to their counter parts. Windows Remote Desktop can pass sound, communicate with local printers, and is much better at speed. If I use VNC compared to RDP I simply do not get near the same experience, RDP makes me feel as if I am onsite where VNC has menu and typing lags. I think an excellent Remote software package would be a great add-on as well.
/me nods.
Perhaps as part of an "official" Red Hat solution include a partnering with NoMachine for their NX remote access client and server software?
I've been using the freenx server on our various K12LTSP servers for several years, with NoMachine's NX Windows and Linux clients for remote access from my home to our school's servers. Especially when connecting over the Internet, I've found the NX Client to have significantly faster response times than VNC. NX Client also supports local file sharing, local printing, and sound... though in full disclosure I've not tried those features.
So I guess to summarize I would love to see the following come about in the future:
- Seamless Cluster Management. With Windows 2003 server I can fire up
a management console and choose server roles, Linux needs this. I would love that when I fire up my new Redhat Server it asked me what roles I would like to assign to this machine. Roles could be Primary or Secondary server, Application server (gui to walk through which apps to export to other servers), DHCP server, File Server, Directory Server, DNS Server, etc. Then when I choose each option it walks me through a little wizard to configure the server role. If I happen to add a server and choose Secondary, it would allow me to pick which primary server to choose for each role with a simple gui that let me enter in the IP address or name of which server was the primary for File serving, Authentication, Applications and so on. I really think a nice GUI server role management tool would go a long way.
- Strong Multimedia delivery. I know there has been a lot of talk that
a move PulseAudio as opposed to ESD or ARTS should help. But along with sound multimedia browser integration and a single do all player would be great. I know multimedia is a must for schools out of the box.
- A good compression remote desktop tool that can transport sound and
recognize local printers.
- RedHat is obviously closely partnered with Dell since RedHat is the
only distro Dell will install on their servers. Work with Dell to package server setups that can handle different sized schools. If I could call Dell and talk to a server rep and let them know that I am looking at deploying a Redhat based thin client setup in a school with 800 terminals and the rep could tell me that I need servers XYZ and ship them as a configured package that I can just plug into my network and start adding terminals, I would be in heaven. This wouldn't go just for schools either, businesses could benefit from this as well. Right now when I mention such a thing to a Dell server Rep they haven't a clue what I am even talking about. There could be a few case studies and some testing to determine how much hardware is needed and what the most efficient way to cluster would be. (Personally I find a speed hit when using Samba/LDAP with NFS mounted /home directories, there has to be a more efficient way of doing things.) Redhat is definitely know as a stable server operating system, couple that with Dell hardware support and preconfigured packages, and you could stumble into a goldmine. Maybe if Redhat and Dell worked together you could find 64bit server setups with quad quad core Zeon processors and 32GB of RAM and 4 teamed Gigabit NICS that could handle 1000 clients from a single machine. People like me just don't have the resources to test this sort of thing and see if it is possible.
- Work with Dell or some other vendors and build an approved client
list that is known to work with the above preconfigured packages out of the box. This could provide a start to finish package that is guaranteed to work. Maybe Dell could even look at getting into the thin client market along with RedHat.
- Be sure that all common Education apps are in RedHat software
repository and can easily be installed with Yum Extender. I don't think that they need to be prepackaged, but need to be easily searched as Education applications and be available. A good scheduling/grading appliation such as http://richtech.ca/openadmin/ or http://www.miller-group.net/index.php?option=com_frontpage&Itemid=1 are a must in this list.
- Possibly build local vendor support for the sale and technical
support of such systems. I know my business would be interested in supporting Southern Minnesota, Wisconsin, and Iowa area. http://www.1-cs.com. Local support options may be a key in implementation and making end users feel comfortable. If nothing else knowing support is available could set minds at ease.
I think if Redhat could find solutions to the above problems they could dominate one to one computing and the terminal services market in both Education and Business. I am not sure how much of the above is even in the realm of RedHat's scope but if nothing else they may be able to be a good facilitator.
And you thought your message was long :-) Jim Kronebusch
Heh. :)
Thanks for the feedback, Jim. I appreciate it.
--g
David Whitmer Director of Media & Technology Calvary Schools of Holland (Michigan) web: www.calvaryschoolsholland.org email: the.whitmers@gmail.com
Perhaps as part of an "official" Red Hat solution include a partnering with NoMachine for their NX remote access client and server software?
I've been using the freenx server on our various K12LTSP servers for several years, with NoMachine's NX Windows and Linux clients for remote access from my home to our school's servers. Especially when connecting over the Internet, I've found the NX Client to have significantly faster response times than VNC. NX Client also supports local file sharing, local printing, and sound... though in full disclosure I've not tried those features.
NX was the first alternative that popped into my mind as well, but I have never used it myself so I can't recommend it personally yet. However I have read many posts stating that it is much faster and supports the other features.
I wasn't aware that Indiana was using remote connections as opposed to laptops. I am encouraged to hear Greg's feedback that the students see this in a positive way. Having to physically carry a laptop or textbooks with you in such a digital age seems silly. I don't own a laptop personally anymore because it is easier to remote connect, why lug stuff around. My approach to this when I was in school was to leave everything at school, but that isn't necessarily a good thing :-) Enabling students to remotely work from school anywhere with an internet connection and without lugging material around seems like a good solution.
I have also been looking into the alternative of digital textbooks. I have requested some information from Apex Learning (www.apexlearning.com) as they seem to be a leader in this area and Firefox is a supported browser (still not very educated in this area as I am just starting to research the possibilities). I really like the idea of subscription based online curriculum, quizzes, and such. Some of the things they have to offer are intriguing (such as personalized goals through online pretests). I am still not sure how this would be accepted by teachers and students (you can't read a book on a bus anymore or while on vacation without internet access, rural students may suffer as well). But having curriculum that is updated yearly (some of our textbooks are 15 years old) and not having to carry physical textbooks is an interesting concept. Not sure if Redhat could play a part in this however. I don't know what costs are associated with such an initiative at this point either, but to make it successful on thin clients, seamless sound delivery and viewing of Flash would be a must. And if there is a possibility of 400 students to be using firefox with flash all day to reference the material, one would definitely need a powerful server cluster and possibly multiple application servers to run firefox. So I guess my point is that if Redhat can provide an effective platform for delivery, going digital will be much easier.
Jim Kronebusch
On 4/16/07, Greg Dekoenigsberg gdk@redhat.com wrote:
What should Red Hat be doing in the education space?
Some more information on Indiana:
They aren't using Linux thin client for a couple of reasons. First, they are purchasing new, Dell and whitebox PCs for under $300, and they get the benefit of thick-client usage. They use Ardence mostly to stream down the full OS to each PC. According to some conversations I've had with Mike, they haven't felt that LTSP would give them the functionality that they want, and didn't see a huge price savings.
However, that may now be changing. One of their most difficult hurdles is that the full PC generate a lot of heat, and they have to upgrade the air conditioning in the school buildings--and that's turning out to be a problem. Using some low-watt thin-clients may be the answer for them.
They are also experiencing some difficulties with Ardence. They were bought by Citrix, which is owned by Microsoft. I've been in contact with Mike from fl_TeacherTool, as he has visited a site in Canada that is running thick-client streaming full FLOSS. Seems like this might be a good option for them as well.
Steve
-- Steve Hargadon steve@hargadon.com www.SteveHargadon.com 916-899-1400
On 4/16/07, Greg Dekoenigsberg gdk@redhat.com wrote:
What should Red Hat be doing in the education space?
Some more information on Indiana:
They aren't using Linux thin client for a couple of reasons. First, they are purchasing new, Dell and whitebox PCs for under $300, and they get the benefit of thick-client usage. They use Ardence mostly to stream down the full OS to each PC. According to some conversations I've had with Mike, they haven't felt that LTSP would give them the functionality that they want, and didn't see a huge price savings.
However, that may now be changing. One of their most difficult hurdles is that the full PC generate a lot of heat, and they have to upgrade the air conditioning in the school buildings--and that's turning out to be a problem. Using some low-watt thin-clients may be the answer for them.
They are also experiencing some difficulties with Ardence. They were bought by Citrix, which is owned by Microsoft. I've been in contact with Mike from fl_TeacherTool, as he has visited a site in Canada that is running thick-client streaming full FLOSS. Seems like this might be a good option for them as well.
Steve
-- Steve Hargadon steve@hargadon.com www.SteveHargadon.com 916-899-1400
On Tue, 17 Apr 2007 12:19:12 -0700, Steve Hargadon wrote
On 4/16/07, Greg Dekoenigsberg gdk@redhat.com wrote:
What should Red Hat be doing in the education space?
Some more information on Indiana:
They aren't using Linux thin client for a couple of reasons. First, they are purchasing new, Dell and whitebox PCs for under $300, and they get the benefit of thick-client usage. They use Ardence mostly to stream down the full OS to each PC. According to some conversations I've had with Mike, they haven't felt that LTSP would give them the functionality that they want, and didn't see a huge price savings.
I agree that some of the shortcoming from thin terminals can more easily be solved with network booted thick clients. But the idea of the lower power usage machines and the longer life cycle of thin clients seems better in the long run, but those few hurdles need to be overcome with multimedia and server clustering to really work well. A thin client mounted on a flat screen seems the most efficient.
From what I understand Indiana isn't necessarily forcing Linux either, the schools have
the choice of putting Windows or Linux on the clients. I don't necessarily disagree with that approach, I think schools can benefit from running Windows, OSX and Linux together to better prepare them for the real world. I see Linux as being the means to distribute mass computing (one to one), but I still see the need for a Windows lab or and OSX lab to teach more specific courses for some schools. That being said, a Linux cluster should also be able to authenticate Windows and OSX clients and provide a roaming profile seamlessly between all 3 platforms.
I also see our teachers and staff at our schools using thick clients, and students using thin clients.
Jim
On Tue, 17 Apr 2007, Steve Hargadon wrote:
They aren't using Linux thin client for a couple of reasons. First, they are purchasing new, Dell and whitebox PCs for under $300, and they get the benefit of thick-client usage. They use Ardence mostly to stream down the full OS to each PC. According to some conversations I've had with Mike, they haven't felt that LTSP would give them the functionality that they want, and didn't see a huge price savings.
However, that may now be changing. One of their most difficult hurdles is that the full PC generate a lot of heat, and they have to upgrade the air conditioning in the school buildings--and that's turning out to be a problem. Using some low-watt thin-clients may be the answer for them.
They are also experiencing some difficulties with Ardence. They were bought by Citrix, which is owned by Microsoft. I've been in contact with Mike from fl_TeacherTool, as he has visited a site in Canada that is running thick-client streaming full FLOSS. Seems like this might be a good option for them as well.
/me hrms.
So it seems like we've got a number of factors to consider, all of which seem to point to different approaches:
* Power consumption / Green-ness. Favors thin client.
* Cost. Favors thin client, but only slightly.
* Functionality. Favors thick client in multimedia situations.
* Red Hat's ability to deliver. Frankly, we're better able to deliver a thick client right now from a technology perspective. Packaging and supporting K12LTSP is not our strength... at least not yet.
--g
On 4/17/07, Greg Dekoenigsberg gdk@redhat.com wrote:
On Tue, 17 Apr 2007, Steve Hargadon wrote:
snip......
- Red Hat's ability to deliver. Frankly, we're better able to deliver a
thick client right now from a technology perspective. Packaging and supporting K12LTSP is not our strength... at least not yet.
The plan for K12LTSP, from its inception, was to have it folded into Red Hat. At one time there was a channel in Red Hat Network for K12LTSP. Eric Harrison has been working for a long time, releasing update after update of K12LTSP based on first RH, then Fedora (as well as RHEL and CENTOS). He's configured packaging to comply with RH standards so that whenever RH decides it wants to make billions by distributing a viable thin-client to the world, LTSP will be there, ready to go. (Don't forget Jim McQuillen's hard work too.)
I think though that Eric is getting tired. You can only volunteer so much time over so many years and then you start to get tired. He's not complaining or saying it, but I see it. He has every right to be.
I have to give credit where it's due. RH helped send me and several groups of high school kids to various conferences to show folks how K12LTSP worked. We setup the first big FOSS/K12LTSP lab at NECC with RH's help. Now the open source events at NECC are so large and so popular that vendors who have to pay for their space are complaining. ;-^) (Way to go Steve!!!)
Does RH support sendmail or do they rely on the sendmail user community to support the product? They should just go ahead and include that "Terminal Server" button in the normal Fedora and RHEL install and go for it. They already host the K12OSN listserv. Put a couple of people on the payroll to help flush out the K12LTSP support wiki and answer questions on the list. Rebrand the product RH DesktopTerminalServer and combine it with thick-client deployment and imaging support and just watch it take off.
There is a HUGE market waiting for the first Linux distributor who can bring a viable desktop to the masses. Red Hat has navigated a clear path AWAY from the desktop so far. They've even gone as far as to distance the desktop product from their branded name by spawning off the Fedora project. Reading tea leaves here but in doing this, they have a solid track record of profitability with their server products and they have avoided making the 800lb. gorilla angry.
We need to have RH grow some cojones http://en.wikipedia.org/wiki/Cojonesand find a way to address the protocols issues (for sound and video over the web) and deliver a product.
BTW, OLPC workstations would sell like hotcakes in US schools at $200 a pop. Every household in the country would be buying them to use at home too. What do you think would happen to the protocol standards if there were millions of clients that were configured to use open standards in homes all over the world?
One final note to the folks at RH... Don't underestimate the value of having a loyal base of users in the education market. It saved Apple from certain ruin during their dark years. You have a chance here to make huge inroads in US and international markets by putting out a solid suite of desktop solutions aimed at the K12 market.
;-) Paul
Does RH support sendmail or do they rely on the sendmail user community to support the product? They should just go ahead and include that "Terminal Server" button in the normal Fedora and RHEL install and go for it. They already host the K12OSN listserv. Put a couple of people on the payroll to help flush out the K12LTSP support wiki and answer questions on the list. Rebrand the product RH DesktopTerminalServer and combine it with thick-client deployment and imaging support and just watch it take off.
There is a HUGE market waiting for the first Linux distributor who can bring a viable desktop to the masses. Red Hat has navigated a clear path AWAY from the desktop so far. They've even gone as far as to distance the desktop product from their branded name by spawning off the Fedora project. Reading tea leaves here but in doing this, they have a solid track record of profitability with their server products and they have avoided making the 800lb. gorilla angry.
We need to have RH grow some cojones http://en.wikipedia.org/wiki/Cojonesand find a way to address the protocols issues (for sound and video over the web) and deliver a product.
BTW, OLPC workstations would sell like hotcakes in US schools at $200 a pop. Every household in the country would be buying them to use at home too. What do you think would happen to the protocol standards if there were millions of clients that were configured to use open standards in homes all over the world?
One final note to the folks at RH... Don't underestimate the value of having a loyal base of users in the education market. It saved Apple from certain ruin during their dark years. You have a chance here to make huge inroads in US and international markets by putting out a solid suite of desktop solutions aimed at the K12 market.
;-) Paul
Well said Paul. I would love to see RedHat adopt LTSP as a base package. It seems a lot of the thin client issues need to be addresses farther upstream at the OS level. Otherwise all other fixes just seem to be hacks and patches. Ubuntu seems to be winning the race at this point for adopting LTSP into their OS and really working to make things work at the OS level instead of being patched downstream. But LTSP 5 still has a lot of room to grow as far as speed and scalability, this is still a great time for Redhat to jump in.
I really wish I could get those OLPC laptops as well, they sure would make one-to-one computing an affordable possibility. I have never liked laptop programs, but those could make it work. The best idea I saw for bringing the OLPC to the US was to charge $200 per machine where as $100 would go towards the hardware purchase, and the other $100 would sponsor a laptop in a third world country. I know our school would love such a program, it would bring them low dollar computing and also give some pride in knowing they others less fortunate were being helped.
On Tuesday 17 April 2007 14:18, Steve Hargadon wrote:
On 4/16/07, Greg Dekoenigsberg gdk@redhat.com wrote:
What should Red Hat be doing in the education space?
Some more information on Indiana:
They aren't using Linux thin client for a couple of reasons. First, they are purchasing new, Dell and whitebox PCs for under $300, and they get the benefit of thick-client usage. They use Ardence mostly to stream down the full OS to each PC. According to some conversations I've had with Mike, they haven't felt that LTSP would give them the functionality that they want, and didn't see a huge price savings.
However, that may now be changing. One of their most difficult hurdles is that the full PC generate a lot of heat, and they have to upgrade the air conditioning in the school buildings--and that's turning out to be a problem. Using some low-watt thin-clients may be the answer for them.
They are also experiencing some difficulties with Ardence. They were bought by Citrix, which is owned by Microsoft. I've been in contact with Mike from fl_TeacherTool, as he has visited a site in Canada that is running thick-client streaming full FLOSS. Seems like this might be a good option for them as well.
I hadn't heard that part. My understanding is that Ardence is basically doing NFS root of a full OS. Which is similar to what LTSP 5 (muekow) is doing on Ubuntu and a few other distributions .
On 4/17/07, Richard June rjune@bravegnuworld.com wrote:
I hadn't heard that part. My understanding is that Ardence is basically doing NFS root of a full OS. Which is similar to what LTSP 5 (muekow) is doing on Ubuntu and a few other distributions
Speaking of ltsp 5.... Are you game for helping me take eric harrison's current ltsp 5 package set and getting them into shape for submission to Fedora Extras? I'm going to be spending time this week and next week poking at what he's currently got with the goal of making a package submission to fedora extras as soon as I am comfortable with the mechanics of the packaging. My understanding from my brief email conversation with eric is that there could be some improvement in automating the configuration process for a default config, but that the packages basically work.
I'll stand up and act as maintainer for the package set for the time being until I find some co-maintainers with cvs access. Once we get ltsp 5 into the fedora repository we can have a serious discussion as part of the fedora development dialog about how we can establish a ltsp install target for fedora 8 and beyond. Post Core/Extras merger we might even have a good chance of getting a dedicated ltsp or even an educational spin out into the wild IF we have enough momentum to keep it maintained. If we can get a fedora spin for ltsp together (that can be tracked with download metrics), we'll be in a strong position to shame Red Hat into putting some engineering manhours into the development to start pulling a product together based on the Fedora packages.
-jef
Yee haw cowboy. lets ride. On Wednesday 18 April 2007 00:21, Jeff Spaleta wrote:
On 4/17/07, Richard June rjune@bravegnuworld.com wrote:
I hadn't heard that part. My understanding is that Ardence is basically doing NFS root of a full OS. Which is similar to what LTSP 5 (muekow) is doing on Ubuntu and a few other distributions
Speaking of ltsp 5.... Are you game for helping me take eric harrison's current ltsp 5 package set and getting them into shape for submission to Fedora Extras? I'm going to be spending time this week and next week poking at what he's currently got with the goal of making a package submission to fedora extras as soon as I am comfortable with the mechanics of the packaging. My understanding from my brief email conversation with eric is that there could be some improvement in automating the configuration process for a default config, but that the packages basically work.
I'll stand up and act as maintainer for the package set for the time being until I find some co-maintainers with cvs access. Once we get ltsp 5 into the fedora repository we can have a serious discussion as part of the fedora development dialog about how we can establish a ltsp install target for fedora 8 and beyond. Post Core/Extras merger we might even have a good chance of getting a dedicated ltsp or even an educational spin out into the wild IF we have enough momentum to keep it maintained. If we can get a fedora spin for ltsp together (that can be tracked with download metrics), we'll be in a strong position to shame Red Hat into putting some engineering manhours into the development to start pulling a product together based on the Fedora packages.
-jef
Fedora-education-list mailing list Fedora-education-list@redhat.com http://www.redhat.com/mailman/listinfo/fedora-education-list
On 4/16/07, Greg Dekoenigsberg gdk@redhat.com wrote:
The question, for those who want to skip to the end:
What should Red Hat be doing in the education space?
But OLPC isn't enough. For one thing, it's designed to solve problems that kids in the developed world don't have, and thus may be regarded by many in the developed world as an expensive toy -- especially when a desktop system can be had for $300 or less. (This is true even in some of OLPC's target markets; you wouldn't believe the resentment I heard from some Brazilians at being considered "too poor for real computers".) For another, it may not work in the classroom nearly as well as we think; if even one or two kids leave their laptops at home, suddenly you don't have a one-to-one classroom anymore. Most of all, though, it's just unwise to put all of your eggs in one basket -- and that is precisely what Red Hat has been doing in the worldwide K12 space with OLPC.
I think Red Hat should put all its eggs in the OLPC basket.
I don't see anything about the XO and Sugar that wouldn't solve problems right here in the south side of Providence, RI. That is, unless OLPC Just Doesn't Work, in which case it will be of no use to anyone, here or abroad.
In particular, Sugar is probably the only chance we get for a decade or more to craft a desktop environment and suite of applications specifically for learning and collaboration. Even if schools don't start using the XO hardware, I'd like to see Red Hat making it easy to use Sugar in schools, on thin clients, etc.
We need to do what we can to make OLPC a success and as a community ride on its coattails into US schools.
--Tom
On 4/17/07, Tom Hoffman tom.hoffman@gmail.com wrote:
I think Red Hat should put all its eggs in the OLPC basket.
I don't see anything about the XO and Sugar that wouldn't solve problems right here in the south side of Providence, RI. That is, unless OLPC Just Doesn't Work, in which case it will be of no use to anyone, here or abroad.
I think there is definitely room to move towards introduction of Sugar into US schools..sure. But I think you need to make some less aggressive targets for near term successes that build inroads into US local school systems at all levels that matter: teachers, IT departments, and school boards; before can can have a constructive serious discussion about Sugar as a new an exciting forward looking technology.
Less aggressive targets would include well-packaged and sustainable moodle deployments and an establishment of a teacher/open-technology development network. You get those in place and you make school systems feel comfortable working with Red Hat as a partner in developing technical solutions and you have a starting point towards focusing a larger community towards Sugar as the new interface for educational computing.
-jef"The Openmind network: Where education finds open technology solutions"spaleta
On 4/17/07, Jeff Spaleta jspaleta@gmail.com wrote:
On 4/17/07, Tom Hoffman tom.hoffman@gmail.com wrote:
I think Red Hat should put all its eggs in the OLPC basket.
I don't see anything about the XO and Sugar that wouldn't solve problems right here in the south side of Providence, RI. That is, unless OLPC Just Doesn't Work, in which case it will be of no use to anyone, here or abroad.
I think there is definitely room to move towards introduction of Sugar into US schools..sure. But I think you need to make some less aggressive targets for near term successes that build inroads into US local school systems at all levels that matter: teachers, IT departments, and school boards; before can can have a constructive serious discussion about Sugar as a new an exciting forward looking technology.
Less aggressive targets would include well-packaged and sustainable moodle deployments and an establishment of a teacher/open-technology development network.
I'm not sure what a "teacher/open-technology development network" looks like, but I think the incremental advances like Moodle deployments are happening already and will continue to happen with or without Red Hat's participation.
What we're missing now is the complimentary push for a big, ambitious vision (that is simultaneously quite practical) that will catch people's attention and make them think of free software as something other than a low-rent alternative. OLPC's the best chance we're going to get for that. It is our big splash. And I think it is the right thing, pedagogically and technologically.
--Tom
What we're missing now is the complimentary push for a big, ambitious vision (that is simultaneously quite practical) that will catch people's attention and make them think of free software as something other than a low-rent alternative. OLPC's the best chance we're going to get for that. It is our big splash. And I think it is the right thing, pedagogically and technologically.
--Tom
Being that this initiative is being developed in the US, I sure wish we weren't put on the back burner. We sure could use the OLPC program in our school, but from what I read that won't be a possibility for a few years.
On Wed, 18 Apr 2007, Jim Kronebusch wrote:
What we're missing now is the complimentary push for a big, ambitious vision (that is simultaneously quite practical) that will catch people's attention and make them think of free software as something other than a low-rent alternative. OLPC's the best chance we're going to get for that. It is our big splash. And I think it is the right thing, pedagogically and technologically.
--Tom
Being that this initiative is being developed in the US, I sure wish we weren't put on the back burner. We sure could use the OLPC program in our school, but from what I read that won't be a possibility for a few years.
If you can convince one of the states in the union to pony up for the cash to guarantee a million units, I'll bet you could be on the front burner pretty quick. Seriously.
--g
On 4/18/07, Greg Dekoenigsberg gdk@redhat.com wrote:
On Wed, 18 Apr 2007, Jim Kronebusch wrote:
What we're missing now is the complimentary push for a big, ambitious vision (that is simultaneously quite practical) that will catch people's attention and make them think of free software as something other than a low-rent alternative. OLPC's the best chance we're going to get for that. It is our big splash. And I think it is the right thing, pedagogically and technologically.
--Tom
Being that this initiative is being developed in the US, I sure wish we weren't put on the back burner. We sure could use the OLPC program in our school, but from what I read that won't be a possibility for a few years.
If you can convince one of the states in the union to pony up for the cash to guarantee a million units, I'll bet you could be on the front burner pretty quick. Seriously.
I just wonder if we couldn't get Sugar onto our existing systems before we try to convince our state governments to buy a million laptops (and Quanta to sell us a million laptops?)? It seems like an easier bar to cross.
--Tom
On Wed, 18 Apr 2007, Tom Hoffman wrote:
I just wonder if we couldn't get Sugar onto our existing systems before we try to convince our state governments to buy a million laptops (and Quanta to sell us a million laptops?)? It seems like an easier bar to cross.
--Tom
I think that Sugar has a-ways to go before it can be considered a panacaea. If you guys think it's legitimately something that educators in the US will go for as a Desktop OS for the classroom, we could certainly think about repackaging it for broader use.
The nice thing about the OLPC, of course, is that the hardware profile is 100% known -- which allows the developers to optimize Sugar for the OLPC. That said, it certainly runs on other platforms -- my laptop running FC6, for instance.
--g
On Wed, 18 Apr 2007 11:46:20 -0400 (EDT), Greg Dekoenigsberg wrote
On Wed, 18 Apr 2007, Tom Hoffman wrote:
I just wonder if we couldn't get Sugar onto our existing systems before we try to convince our state governments to buy a million laptops (and Quanta to sell us a million laptops?)? It seems like an easier bar to cross.
--Tom
I think that Sugar has a-ways to go before it can be considered a panacaea. If you guys think it's legitimately something that educators in the US will go for as a Desktop OS for the classroom, we could certainly think about repackaging it for broader use.
The nice thing about the OLPC, of course, is that the hardware profile is 100% known -- which allows the developers to optimize Sugar for the OLPC. That said, it certainly runs on other platforms -- my laptop running FC6, for instance.
I am not too sure the best platform is Sugar. I am not the most educated as to the workings of sugar, so maybe I am talking out of my realm here. But Sugar looked to be prepackaged with certain apps and functionality. I would still want to be able to customize my application offerings for deployment instead of using a predetermined application set and I may also want to tweak the interface to fit our users needs. But maybe sugar isn't as flat as it looked to me at first.
On 4/18/07, Jim Kronebusch jim@winonacotter.org wrote:
I am not too sure the best platform is Sugar. I am not the most educated as to the workings of sugar, so maybe I am talking out of my realm here. But Sugar looked to be prepackaged with certain apps and functionality. I would still want to be able to customize my application offerings for deployment instead of using a predetermined application set and I may also want to tweak the interface to fit our users needs. But maybe sugar isn't as flat as it looked to me at first.
You could think of Sugar as a desktop environment like KDE or GNOME (architecturally, it is sort of a cousin to GNOME). There is more to it than that, but it'll do for starters.
It should be possible (in the future) to allow a student to chose whether to log into GNOME or Sugar on a LTSP terminal. It also should be possible to allow them to run Sugar as an application within their regular GNOME session, in full screen or windowed mode.
Sugar is extensible through its own packaging system.
And it is certainly not flat; if it fails, it will be because they're trying to go too deep.
--Tom
You could think of Sugar as a desktop environment like KDE or GNOME (architecturally, it is sort of a cousin to GNOME). There is more to it than that, but it'll do for starters.
It should be possible (in the future) to allow a student to chose whether to log into GNOME or Sugar on a LTSP terminal. It also should be possible to allow them to run Sugar as an application within their regular GNOME session, in full screen or windowed mode.
Sugar is extensible through its own packaging system.
And it is certainly not flat; if it fails, it will be because they're trying to go too deep.
Thanks for the explanation, I thought of Sugar more as an OS and not just a window manager. Very cool. Tom, maybe you could get Indiana on board and bring OLPC to the US :-)
If you can convince one of the states in the union to pony up for the cash to guarantee a million units, I'll bet you could be on the front burner pretty quick. Seriously.
Hmmm :-)
On 4/18/07, Greg Dekoenigsberg gdk@redhat.com wrote:
If you can convince one of the states in the union to pony up for the cash to guarantee a million units, I'll bet you could be on the front burner pretty quick. Seriously.
A million units in alaska...lets see thats like 10 laptops per child.. you'll need to change the name of the project slightly.
-jef
education@lists.fedoraproject.org