Re: Notes from a Planning Session
(APs in infrastructure mode) is what is happening on the ground mostly. Still, it's not good as it kills the mesh-to-the-school scheme which is one of the key technical goals. This should be easy to fix. Re-enable the NetworkManager code that allows any laptop to gateway the mesh to the Internet. This all used to work in pre-B4 software, at least for USB Ethernet interfaces. As long as the wifi chip can simultaneously do meshing and register with an access point (eth0 and msh0) then Internet access over a WiFi access point should look the same to NetworkManager as a wired eth0. Each laptop near the school will use the AP; any further away would mesh to a laptop near the school. John ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: Notes from a Planning Session
On Fri, Apr 11, 2008 at 8:28 AM, John Gilmore [EMAIL PROTECTED] wrote: (APs in infrastructure mode) is what is happening on the ground mostly. Still, it's not good as it kills the mesh-to-the-school scheme which is one of the key technical goals. This should be easy to fix. Re-enable the NetworkManager code that allows any laptop to gateway the mesh to the Internet. This all used to work in pre-B4 software, at least for USB Ethernet interfaces. As long as the wifi chip can simultaneously do meshing and register with an access point (eth0 and msh0) then Internet access over a WiFi access point should look the same to NetworkManager as a wired eth0. IIRC the problem with this is that collaboration fails on the XO acting as the gateway. See #3136, #2980 - and of course your own #4641. Morgan ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: Notes from a Planning Session
2008/4/11 Carol Lerche [EMAIL PROTECTED]: Thank you for sharing this discussion. Upon reading it I had two questions. Sugar. I have seen offers on this list from a class ofuniversity graduate students to do usability testing. Maybe someone responded to them privately. (That would have been perfectly appropriate.) But in reading the portion of the planning discussion about APIs and sugar, I was struck by the unstated assumption that the sugar interface is unquestionably a good thing. Where is the usability testing with children of the age groups OLPC targets that proves that this is so, in comparison with a more conventional desktop model? In watching 5-6 year olds use the interface for a week, I was struck by sugar's complexity in pursuit of simplicity. It was a difficult interface for the children to learn. Too many steps, including going among different screens. Perhaps I am wrong. But I would like to see the same care in usability testing for younger kids that has been given to ensuring that all the underlying components are written in Python and thus potentially modifiable by the oldest target audience. Please point me to the usability studies I have missed. http://www.gnome.org/~seth/blog/onusabilitytesting Marco ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: Notes from a Planning Session
Thank you for sharing this discussion. Upon reading it I had two questions. *Sugar.* I have seen offers on this list from a class ofuniversity graduate students to do usability testing. Maybe someone responded to them privately. (That would have been perfectly appropriate.) But in reading the portion of the planning discussion about APIs and sugar, I was struck by the unstated assumption that the sugar interface is unquestionably a good thing. Where is the usability testing with children of the age groups OLPC targets that proves that this is so, in comparison with a more conventional desktop model? In watching 5-6 year olds use the interface for a week, I was struck by sugar's complexity in pursuit of simplicity. It was a difficult interface for the children to learn. Too many steps, including going among different screens. Perhaps I am wrong. But I would like to see the same care in usability testing for younger kids that has been given to ensuring that all the underlying components are written in Python and thus potentially modifiable by the oldest target audience. Please point me to the usability studies I have missed. *Networking:* how much of the problems still outstanding with the mesh could be addressed by throwing money at it, i.e. sending commercial APs to the small schools not currently using them. If this addresses the most pressing problems in the most common use cases, wouldn't it make sense to do so, given the opportunity cost of devoting engineering resources to the mesh issues when so many other problems vie for this time? Carol Lerche On Thu, Apr 10, 2008 at 6:18 PM, Michael Stone [EMAIL PROTECTED] wrote: Dear devel@ cjb, marcopg, tomeu, myself, and several others conducted a 2-hour planning session this morning. I've created a transcript of that discussion [1]. If you're interested, please review the questions that were raised and contribute your thoughts. (FYI: The end goal of this effort is a convincing written statement of where we want to go in the next four months, why we want to go there, and how we intend to get there.) Michael [1]: http://wiki.laptop.org/go/User:Mstone/August_planning ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel -- Always do right, said Mark Twain. This will gratify some people and astonish the rest. ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: Notes from a Planning Session
I second Carol's pragmatic approach. What we should do is to use access points in schools whenever possible. The mesh network was not designed to compete with infra-structure. It was designed to be used when there is no infra-structure. That being said, I will keep repeating myself that non-infrastructure is the default and will be the default scenario for years to come. So I believe we want and we need to have a working mesh and the good news is: we have! The kids under the tree scenario (i.e. the mesh) works and it works pretty well. In my testbed with 10 XOs removed from any infra-structure, the XOs can effectively collaborate. What we have are scalability issues that are being addressed gradually. And application issues that will happen even in infrastructure mode. The fact that we can't have 50 XOs collaborating through the mesh now does not mean we don't have a working mesh. And it is important for all the involved people to realize that we will progress to have N laptops, but there'll always be N+1. *Networking:* how much of the problems still outstanding with the mesh could be addressed by throwing money at it, i.e. sending commercial APs to the small schools not currently using them. If this addresses the most pressing problems in the most common use cases, wouldn't it make sense to do so, given the opportunity cost of devoting engineering resources to the mesh issues when so many other problems vie for this time? Carol Lerche On Thu, Apr 10, 2008 at 6:18 PM, Michael Stone [EMAIL PROTECTED] wrote: Dear devel@ cjb, marcopg, tomeu, myself, and several others conducted a 2-hour planning session this morning. I've created a transcript of that discussion [1]. If you're interested, please review the questions that were raised and contribute your thoughts. (FYI: The end goal of this effort is a convincing written statement of where we want to go in the next four months, why we want to go there, and how we intend to get there.) Michael [1]: http://wiki.laptop.org/go/User:Mstone/August_planning ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel -- Always do right, said Mark Twain. This will gratify some people and astonish the rest. ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: Notes from a Planning Session
2008/4/10 Ricardo Carrano [EMAIL PROTECTED]: I second Carol's pragmatic approach. What we should do is to use access points in schools whenever possible. The mesh network was not designed to compete with infra-structure. It was designed to be used when there is no infra-structure. Don't worry - this (APs in infrastructure mode) is what is happening on the ground mostly. Still, it's not good as it kills the mesh-to-the-school scheme which is one of the key technical goals. The kids under the tree scenario (i.e. the mesh) Yes - that works, but it's not the whole picture. cheers, m -- [EMAIL PROTECTED] [EMAIL PROTECTED] -- School Server Architect - ask interesting questions - don't get distracted with shiny stuff - working code first - http://wiki.laptop.org/go/User:Martinlanghoff ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
