[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]

Re: [school-discuss] Educational Killer App for OpenSource

Engineering problems indeed! I think the wireless bandwidth problem is going away soon with 802.11n, even for 25-30 in a classroom and 1:1 netbooks. I only see a max of about 3 Mbps per client when running highly animated applications (elementary edutainment or games) or say video on a web site, so for 30 students, that's only 90 Mbps, well in the reach of 802.11n at 5 GHz (multichannel), 300 Mbps, and maybe usable with 2.4 GHz version (130 Mbps). Real throughput will be less of course, but definitely doable with 5 GHz 802.11n I'd say.

Yah, the OS load over wireless, that's the real issue for netbooks, but I'm not saying they have to be true thin clients either. I think the wireless netbooks should work as standalone systems for home use, but be slavable to the school network (and teacher's PC) when at school, and enable TeacherTool/iTalc type use cases to obviate the need for interactive whiteboards if possible (not projectors, I still think classrooms need those). Whether they're true thin clients doesn't matter to me as much as the ability for the school to keep viruses off their network and ensure that standard apps like OpenOffice, filtered Internet, preconfigured bookmarks and templates, printing, and student file access anywhere in the school are ubiquitous and seamless. That's what I'd love to give my teachers.


Peter Scheie wrote:
We're getting there, Daniel, but I still see a potential bottleneck: wireless bandwidth. The folks at RevolutionLinux have taken Netbooks and installed a stripped down Linux that just loads the NX client which then connects to the LTSP server. It's not quite thin client, but it's close. It addresses the bandwidth issue that wireless has, but at the expense of needing to install a local OS, however small, and giving up the local application option. Unless you give up local apps--which means video--I don't think wireless has the bandwidth yet. I have wondered what one could do with multiple WAPs in a room but even then I think you'd have more demand than could be supported with 20+ netbooks in a room.

Another pitfall of the above plan is that the netbook is no longer a thin client. Because of the way wireless works--needing to specify parameters such as which SSID to join--I don't know that a netbook that loads its OS via the network is possible. However, I plan to explore putting the OS-NX onto a pico USB stick, and booting from that. They're cheap, about about $6 a piece right now. That way, the techs can just crank out a whole bunch of the sticks, put them into the netbooks, and give some extras to the teachers. When a student gets one that doesn't boot because the previous user managed to trash it (intentionally or otherwise), the teacher can just replace it with one of the extras. Using USB sticks means the techs can prepare them without having to gather up all the netbooks, reducing their load.


Daniel Howard wrote:
After watching teachers and students use the K12LTSP system here in Atlanta, in detail in elementary and now middle school, my vote for the killer app/use case to make OpenSource become truly mainstream in education is...drum roll...

Netbooks capable of also running as thin clients over wireless LAN

Here's why: netbooks, especially when they get below $200, are designed for the key applications for education: web access, office suite (mainly word processing, presentation, and spreadsheet), and to a lesser extent, image manipulation. We set up a MAC lab in my daughter's middle school and while Garage Band, iVideo, etc. were big hits when a teacher was dedicated to running the lab, this year that teacher is doing a science classroom, so guess what the other teachers use the Mac lab for? Web access and word processing.

The only real thing holding netbooks back is the same thing holding back our initial thin client rollout, that of supporting video w/o choppy playout. But if the netbook has at least 250 MB RAM and can run the browser/decoder locally, that would solve that problem, and the far lower electricity requirements coupled with mobility and ability to run them as thin clients over a wireless network, that solves all the classroom wiring issues. And if HiVision ever makes good its claim to offer $98 netbooks, then a netbook would be on the order of the cost of a plotting calculator. For a classroom of say 30 kids, $100 netbooks and a classroom LTSP server would cost about the same as some of those expensive interactive whiteboards, less if the parents supply the netbooks like they often currently do the plotting calculators.

Which is why I want wireless thin client use as well: that opens up all the use cases that interactive whiteboards using applications like TeacherTool and iTalc would allow. Hence my vote for netbooks that can be used as wireless thin clients in addition to their normal standalone use at home. Heck, even the Economist recommends Linux on netbooks now:


Best, Daniel

Evan Leibovitch wrote:
Les R wrote:
Actually, Moodle is probably the killer app in this context, but really I am talking about a need that only Open Source Software and Open Content can fill -
bringing affordable classes to kids.

I'm not so sure that this goal is best brought about by seeking to define a "killer app" anymore. The software world is so competitive and complex that it's hard to think that any single application would be compelling enough to make people change their operating systems and, indeed, their overall approach to computing.

Daniel Howard
President and CEO
Georgia Open Source Education Foundation