[03:40] hi all [16:54] hi [16:59] hi [16:59] joerg: have you been working on the portal project you mentioned some time ago? [17:00] sure :) [17:00] hi [17:01] vmlintu, I am just working on it right now.... the thing is, I somehow need more developers and more motivation. [17:01] a project like that is quite hard in the beginning. [17:02] I have written some specs, you know - but they don't attract anybody.... [17:02] I guess people want a demo / see some code. [17:02] Development is the best motivation for the open source world :) [17:03] joerg: www.myserv-project.org ? [17:03] yes [17:03] I've seen many, many many specs in #edubuntu the last years - only 1 or 2 of them got implemented.. [17:03] hehe [17:03] alkisg, I am developing for some time now, but it is really hard [17:04] if I had an active community around it, I'd be really motivated and be twice as fast I guess. [17:05] As sbalneav says, each one scratches his own back... e.g. I'm developing "school-scripts" because I need them, to use them in my classroom [17:05] So even if I had to do it on my own, I'd still do it [17:05] I need people telling me that they need my work, people who challenge me with new things etc. [17:05] Did you send an announcment on the mailing list? [17:05] yes......that's different [17:05] I am doing that because I have a vision [17:06] I am not a teacher or a student [17:06] I am just enthusiast :P [17:06] If you did, and people didn't respond, are you sure that your project fits some existing need? [17:06] joerg: I've checked the site a couple of times, but as it uses python, I really cannot say much.. [17:07] alkisg, well, ages ago I posted a rough draft..... [17:07] and people like vmlintu responded :) [17:07] but at that stage it was still too early [17:08] vmlintu, you should learn python :) not only to contribute to that project ;) but to make your daily life easier. [17:08] I'm more into ruby [17:12] joerg: We've been working on a new version of the ldap based user management tool we've been using for schools [17:12] vmlintu, is that web based [17:12] ? [17:12] joerg: yep [17:13] joerg: hopefully we get it out before lucid [17:13] that is php based? [17:13] ruby on rails [17:14] good [17:14] that should mean it has a clear MVC structure and is easy to extend ;) [17:15] in contrast to these php things which do great things but mostly have bad code [17:15] we have an old perl catalyst based tool for user management that is now replaced with this [17:16] We be using it to handle some hundred schools with 25-30k users [17:16] we'll [17:18] by default it'll be using openldap + mit kerberos as databases [17:19] is there a demo/preview of it? [17:19] there's no public demo yet [17:23] vmlintu, ok....anyway, I am quite sure that we could integrate it in myserv somehow.... [17:24] if we extend it to authenticate using CAS, we could load it in an iframe and we'd be done [17:25] that's what I am doing with roundcube webmail at the moment [17:25] we are now running rubycas-server with kerberos to provide cas [17:27] we've been doing some testing with oauth to provide rest+oauth api for external applications to fetch user information [17:28] vmlintu, why do external apps need to talk to your user management tool? :P [17:30] There are some educational applications running in the web that need user information like class level, pupil/teacher, groups.. now they need to import the data using files or setup ldap connections that require firewall changes.. [17:30] opensocial is the way to go :P [17:30] opensocial uses rest+oauth :) [17:30] yes [17:31] implementing opensocial is just 100x more work than simple rest+oauth api [17:31] no, it is not. [17:31] it is 100x less work [17:31] ? [17:31] rest+oauth was like 30 lines of code with ror [17:32] because the os rest api is well documented and tested - and there are lots of apps already implementing it. [17:32] well, and? [17:32] don't reinvent the wheel :P [17:33] use the structure/methods that opensocial uses [17:33] in your rest api [17:33] choose the methods (e.g. to retrieve user/group info) and implement them in your rest api [17:34] you don't have to fully implement open social [17:34] that's what we pretty much did [17:34] yeah, then it is fine. [17:34] I just wouldn't call it opensocial unless it's the whole thing [17:34] ok [17:35] the user management tool won't implement any of these, anyway [17:35] and I just wanted to say: don't invent another api :P [17:35] If we do any of that, it'll be a separate application [17:36] google's using saml and that's way more requested feature than opensocial [17:37] sure [17:37] but I am not sure if it takes care of all the other nice things [17:38] like user/app data [17:38] friendships / pictures etc. [17:39] myserv is using it to provide a way to integrate gadgets [17:39] friendships are not on our agenda really [17:39] using javascript / iframes [17:39] he he [17:39] well, I have implemented friendships [17:40] because I think it can be helpful [17:40] if you can share files and/or collaborate with a group of people you trust. [17:42] we are focusing on the k12 segment, so especially the grades 0-6 don't really benefit from that [17:43] we have been thinking about "pre-defined friendships" that would be defined by teachers instead of pupils [17:44] harassment by excluding people from group is a real problem in schools, so the tools need to address that [17:49] hmm [17:49] Or maybe a facebook application could be used to take care of all that.. Everyone's using it anyway, so bringing information from school to it would be probably easiest.. [17:49] you don't have to be friends with everybody, right? :) [17:49] maybe finland is a bit more socialistic :) [17:50] in class setting everyone should get along with others, I think [17:51] yes [17:51] so there is a group called "class 5a" [17:52] but if lisa is closer friends with alice and linda it is perfectly okay, isn't it? [17:53] I've talked about that here and the opinion seems to be that in that case they are most probably already using facebook or something else for their own stuff [17:53] and nobody is going to trust a school-run system for their personal stuff [17:54] but like I said, we don't have real plans for this.. [17:54] why not? [17:54] my idea in the long run is to replace facebook and all that stuff [17:54] you mean trust? [17:55] students have an account at their school [17:55] they need it anyway [17:55] they usually have their friends at the school [17:55] why can't they share (private) stuff there? [17:55] e.g. fotos [17:55] not all their friends are in the same school [17:56] I want ppl to say: well, I have a great tool at school, why do I need facebook? :) [17:56] that's true [17:57] but for those who are I want to be better than facebook [17:57] it'd most probably work for school related stuff, but I really have doubts about private stuff [17:57] (actually here facebook is not dominating, but an older local competitor) [17:58] because I think if a web portal (run by the school) get's a social meeting point for the students, it means that they identify a lot more with school [17:59] that thinking is loosing ground over here - the new thinking is that everybody should be able to use their own tools and the school's system should be able to collect the information from all over the net [18:00] meaning that if someone wants to use blogger for bloggin and someone else uses facebook for blogging, it should work [18:01] personal learning environments are something in that direction [18:04] lool [18:04] a lesson has 45 minutes here [18:05] if the project is to write a blog about something and people are using 5-10 different tools [18:05] and the teacher only knows one of them [18:05] that means 2/3 of the time will be wasted on technical stuff [18:06] instead of producing content. [18:06] we have had that issue a lot of times. [18:06] it doesn't work. [18:06] that's why most of the schools tell their students and teachers: [18:07] we have ONE tool to make a presentation and that is called openoffice.org impress [18:07] i agree. setting the tools for an organization is the only way to go [18:07] and if people don't have it at home, they can get a copy of it at school [18:07] on their flash drive [18:07] and if they prefer powerpoint [18:07] they can use it at home [18:08] the idea behind ple's is that the platform aggregates the content using rss/atom/whatever.. [18:08] and it is their responsability that it runs with openoffice [18:08] yes, I understand. [18:08] maybe in finland things are different. [18:08] joerg: where is that? [18:09] but here, teachers are happy if they more or less know ONE tool [18:09] there is no money for system administration and maintenance [18:09] and teachers are not educated in IT concepts [18:09] alkisg, germany [18:10] Nice, I didn't know openoffice was so widespread there... [18:11] alkisg, I don't say it is wide spread [18:12] it is up to the school [18:13] I am freelancing for an organization that takes care of the IT infrastructure of 200 schools here [18:13] the townships usually have contracts with us [18:13] and they pay a certain amount per student/per year to us [18:14] of course schools can buy a MS office license from microsoft [18:14] that's up to them. [18:15] but we don't spend a single cent on proprietary software we provide to the schools [18:15] open office is good enough. [18:15] and everybody can have it [18:15] no matter if daddy is a rich manager or a poor worker :P [18:17] and well, the PLE stuff is nice [18:17] but that would mean students and teachers (!) would need to have an understanding of the technologies. [18:25] ok, I'm back.. [18:26] over here quite a few schools are using moodle, but not all.. some use blogger or other blogging services and some use other learning environments [18:27] and I've been hearing people talking about mahara lately [18:27] that means that if we provide schools a portal, it'll be definitely different from the tools they are already using and the next school won't agree with the tools anyway [18:28] that's why we are not doing any end-user services, but only administration tools and backend services for the actualy tools used by the end-users [18:30] an aggregator that would collect content from different tools could be a solution here [18:31] So basically the solution would be that the school has a default set of tools (for example moodle + blogger + teachertube) that is taught, but you can use other tools if you know them already [18:33] hehe [18:33] schools here never heard of moodle mostly :P [18:33] apart from that, moodle speaks CAS and can use an ldap db [18:33] so wouldn't be a problem to integrate that in a myserv portal [18:36] moodle is not exactly that great tool, imho [18:37] it's better than the old competitors that it replaced, but it's way too teacher-centric [18:40] most of that stuff is simply overkill for our schools [18:40] they want to upload files somewhere [18:40] and they want to write e-mails [18:41] here some are starting to use dropbox to upload files [18:41] and the very advanced users use a wiki to collect results [18:41] one out of 50 teachers :D [18:41] you can do that [18:41] if you have enough bandwidth [18:42] what kind of connections the schools have there? [18:42] with an ISDN line or a slow DSL you know what happens [18:42] if 25 users upload their documents [18:42] you have to ask people to start uploading 30 minutes before the lesson ends ;) [18:43] some still have ISDN (128k up/dn) [18:43] that can be a problem then.. I thought that 2/1 adsl is bad.. [18:43] DSL 2048k dn/196k up [18:44] schools having 512k up are very lucky [18:46] here many of the schools have 100Mbps fiber to city network that has 10 or 20Mbps internet connection [18:46] some have 100Mbps internet [18:47] it seems like the background is quite different - we have tons of existing tools that should be supported and more bandwidth - you have no existing tools and less bandwidth [18:47] there are already schools that are using google apps for all their classroom work.. [19:26] vmlintu, anyway....we more or less have the same concept [19:26] but we focus on both local and external services [19:26] yes, the basic concept is pretty much the same [19:26] I see the myserv portal as the integrator: it provides a CAS auth server, users can create a profile there, users can found groups [19:27] and in their user and groups profiles they can do "things" [19:27] like adding one of the apps you mentioned [19:27] like an externally hosted blog [19:27] or wiki [19:28] or a vocabulary trainer [19:28] whatever. [19:28] who creates the user accounts? [19:28] the only important thing is that the app "speaks" oauth/cas and opensocial [19:29] if it wants to retrieve information about anything [19:29] the admins do [19:30] django has a nice admin interface [19:30] if the users are in ldap, we don't have a solution yet. [19:30] they would need to use whatever ldap user management tool [19:31] a primary school won't need ldap [19:32] that's why I don't want to marry ldap :P but only provide is as an option. [19:32] we are using ldap for every single school [19:32] lol [19:32] we have primary schools with ten user accounts here [19:32] for what reason should they run ldap? [19:32] setting up a system without ldap takes way more time than with it [19:33] as all schools use the same user management tool, we can support them better === joerg__ is now known as joerg [19:45] sorry [19:46] vmlintu, we don't have resources to support schools managing users :P [19:46] and if they have ten users and only use that web app [19:47] we won't run and maintain an ldap installation for them [19:47] it is just another thing to maintain and keep running [19:47] and another DB to backup [19:48] and the advantage is: none [19:48] we are actually taking care of all the server management, so the schools don't even have to know that they have ldap.. [19:49] yeah [19:49] exactly [19:49] and training for user management tools is part of the service.. and helpdesk they can call [19:50] they won't know it [19:50] so there is no need to do that additional work [19:50] will not be of any use for them. [19:50] the web portal has user and group management in the postgres DB [19:50] what additional work do you mean? [19:50] that's perfectly ok for a system with ten users and no additional services [19:51] additional work means: setting up slapd, running slapd, backing up slapd [19:52] yes, in some cases ldap does not bring any features that would be needed, but in our case running some systems without ldap would make everything much harder [19:52] hmm [19:52] Now we have tons of identical servers that can be managed exactly the same way [19:53] and slapd installation is fully automated with everything else [19:53] and we will have 10-20 servers with a big ldap DB [19:53] and 130 servers with just the webapp [19:53] and a DB to backup [19:54] we are using the same backup system for all the servers with identical settings [19:54] probably these 130 servers will just be one or two big root servers in some data center [19:55] and now that I remember, we need to be running kerberos anyway for the desktop, so ldap comes with it [19:56] finland must be a rich country :P [19:56] how many schools do you support? [19:57] over 100 now [19:57] ok. how many full time jobs do you have for that? [19:58] setting up a server with automatic tools takes about 30 minutes now.. with ldap+kerberos+web tools+desktop+ltsp [19:59] ok [19:59] mostly you just enter the domain name and a root password :D [20:00] we install also the networks and thin clients and give training to teachers and pupils [20:00] just tell me how many full time jobs you have to support these 100 schools. [20:01] 5 [20:01] you see [20:01] we have 200 schools and two. [20:02] actually 150 over years [20:02] what I'm trying to say is that if some of the servers didn't have ldap, we'd need one or two more to support all.. [20:02] but now there's a new contract with some townships [20:02] and we will have 200 very soon [20:02] because that's the only way to hold the 2nd full time job. [20:03] what are the thing you charge for? [20:03] things [20:04] I've understood that you don't provide the desktop itself? [20:04] put a hard disk image on and old computer [20:04] driving there [20:04] setup the "server" there [20:04] driving away and hoping that the hard disk will never fail :) [20:05] it's basically just setting up a filtered proxy, mail, file and print server [20:05] and taking that machine and bring it there. [20:06] no clients :P [20:06] ok, that's quite different from what we do.. we don't provide mail and pretty much nobody wants filtering [20:06] lol? [20:06] linux thin client desktops are the main thing [20:06] so people are allowed to download porn and so on? :D [20:08] we MUST filter the internet, because law requires that. [20:08] I think the reason is that all the filters are so bad that they just encourage the kids to find ways around them.. [20:08] absolutely true [20:09] we use squidguard with blacklists [20:09] it is not that bad [20:09] all the blacklists we have tried also give a lot of false positives [20:10] why? [20:10] teaching russian is not fun when half of the sites are blocked.. [20:10] hmm [20:10] a lot of video and image sites are also listed [20:10] nobody teaches russian here ^^ [20:10] we need to support russian keyboards on most systems.. [20:11] didn't know that there's such a deep "friendship" between finnish and russians ^^ [20:12] I thought you hate them [20:12] and one of the worst filters filtered all the sites that had "sex" in the name.. it just means number 6 in swedish.. [20:14] I know [20:14] that's why we just use these blacklists and that's it [20:15] we are not filtering for any keywords or so [20:17] we have now more or less working blacklist for the ones who want it, but most schools have unfiltered access [20:18] btw, have you checked the tools like elgg or buddypress? [20:19] well [20:19] elgg yes [20:20] didn't like the code [20:20] didn't know how to make that run opensocial apps [20:20] auth with CAS [20:20] and stuff [20:24] these tools all don't provide that pluggable concept [20:24] the idea of a big integrator portal [20:26] I want to be able to say: today we are running our photo blogging app locally, tomorrow we have enough bandwidth and it will be hosted externally without the users noticing anything. [20:26] these tools are all very nice as a standalone solution. [20:27] what's your plan, do you plan on supporting multiple schools/organisations from a single installation? [20:28] but they don't provide the flexiblity of being the portal that is like a bolt tying all these already existing apps together [20:28] yes, that might happen in the future. [20:29] but not for now. [20:29] ok [20:29] but I could imagine that there's a library app for example hosted centrally [20:29] there are a lot of good open source opac / library tools [20:30] that could be on a central server [20:30] only one instance [20:30] I mean the portal - if you are going to run single instance to run the different portals for 100 schools if the network allows it [20:30] using oauth/cas they'd be logged in automatically [20:30] and could manage their books there [20:30] thinking that's "their" library [20:31] but in fact, they are just feeding a big catalogue. [20:31] yes, as I said. [20:31] we gonna do that [20:31] as long as the bandwidth is so bad that we have to run local servers. [20:32] if it is getting better the school will simply have a VPN router and that's it. [20:39] ok, it sounds like the "portal" part that connects the applications is the same for both.. [20:39] user management is different [20:41] providing any end-user services is not in our interest, so having the other portal features is not something we need