# radar # a hackcalendar !! We are in ircs://irc.hackint.org/#radar !! See also (wild collection of events+calendars): https://pads.ccc.de/radar-sources Add yours! --- # Summary The main difference between this project and existing approaches is that we're trying to scrape "everything" - instead of hoping that people would be entering things by themselves. With hackers it cannot work any other way: 1. The world is too big 1. We are too lazy 1. Everyone wants to use own, decentralized system ---- # Intro ---- ### Problem 1. no calendar with proper filter/category/tag system 2. no calendar that isn't somewhat based on others actively entering 3. no source for "all" events, not even for reasonable amount of "some" of them. ---- ### Solution: 1. radar is there https://radar.squat.net 2. radar can import others 3. radar could get a "central" source aggregating all the decentral ones Gitlab: https://0xacab.org/radar ---- # Scraping vs. Trust The auth/trust system with radar.squat.net is beautiful, and actually we should also use this **for trusted spaces and meetings** - but (by now!) I just wanna look on the **events that are "free & open knowledge"** (which doesnt mean "for free" and even includes "commercial" ones). ---- Exactly **like a radar**: A radar has a regular look around, checking the area it can reach out to in it's rounds and rounds, and then - beep! - there is something that showed up. "Users" of the calendar can follow that beep/ping, that has showed up by their filters. **Measuring trust**/quality/authentication on those? Everyone needs to do **by oneself**. (for now?) ---- Problem: * What about event organizers who _don't_ like to have their event showing up in such a central source? * What about providing such a central hub, maybe not good at all? * Trust vs. Srape, as mentioned above. * There are more... ---- => **We'll use radar as frontend, but change the way stuff gets entered/collected?** Could also use agendadulibre.fr as frontend instead. --- # Entering events 0. search the web for sources 1. scrape those calenders 2. add some intelligence 3. import them to radar ---- ## 0. search the web for sources => someone needs to... * ...collect possible calenders that can be scraped * ...find out what exports they'd be serving * ...collect metadata of the calendar/event itself * ...check/define categories of each calender ### Next steps: Do that on http://pads.ccc.de/radar-sources (sva is on it, too) ---- ### Examples: http://www.guckes.net/cal/sources.html https://c3voc.de/eventkalender http://foss.events/ http://hackercalender.org https://lwn.net/Calendar/ https://www.agendadulibre.org ---- ## 1. scrape those calenders ical / spaceapi / atom / cdav / etc => someone needs to... * ...collect what different formats are existing * ...what data those formats would be serving * ...look into those how they could be scraped * ...look into those how radar could import them ---- ## 2. add some intelligence No more intelligence needed than e.g. a spamfilter. This is imho the most difficult part, and still very unclear / work in progress. We first need to check on (1.) before we can make final decisions on this one. ---- We can mark the entries with - "just scraped from calendar xyz.tdl" - "...and briefly edited by human" - "...edited by human who had studies website / has been there / knows the people" - and similar. ---- Also giving them lots of - for us relevant - metadata like - "no/commercial/gov/non-com sponsoring" - "mainly/fully volunteer run" - Governance attendance/speaker (marked/ not marked) - Military attendance/speaker (marked/ not marked) - parking/acco/food situation etc ---- Learning ability: Ideally it pings humans about its assumptions, and they can add/correct (meta)data, or just approve. Based on such experience the thing(tm) gets better and better. Its only heuristics, had discussed this with a guy deeply into learning AIs. It doesn't need to be an AI-thing. ---- ## 3. import them to radar Yeah, so who is up for explaining us what should be the end-format that radar can import? So that its just usable like radar.squat.net, but without having ppl entering stuff, but using the scraped data? Addtion, 18th May: "Categories" and "tags" need to be defined. See session from easterhegg below. --- # GTD Hope possible tasks for everyone are clear from the above?! Also pls reply if you agree/disagree on this proposal! ---- # Even one more thing... I also would like to discuss the possible "bad outcomes" of creating such a calender. But that I'll skip that for now... ---- # ==== OLD STUFF === Needs to be sorted... ## Notes from Meet at Easterhegg ### Radar * multiple languages possible, multiple timezones as well * based on the concept of groups and locations, as authentification layers * maximum of one year - repeating events * => possible feature to have it for longer as 1 year * sorting / filter * category (predefined "globally") * tags (open) * price * wordpress plugin for displaying from radar into wordpress eg all the events of my group * it can import and export ical * radar "should" be able to handle any possible item coming from ical * => possible feature * a group has admins: new user creates an account, the admin can add you then, you're supossed to be part of a group as a user that can enter ## MISC * Why predefine categories? Is community-built or archive-based tag cloud possible? Yes :-) * restricted by group or location? maybe not useful, because of filter-bubble risk of tag terminology * Features: * prices: currency based on location? * submenu of first currency and then amount * automatic tag suggestions (spam status for email) possible? * train filter on previous events? * location drop-down (atm squat yes/no): * hackerspace * bar * office / coworking space * eventlocation * negative filtering / blacklisting * => possible feature * how regular / frequency / extraodinary-ness -> metadata? * frab/ other calender systems exports? not reliable, because might be online only adjecent to event => less planning & archival => ETOOMUCH. knowing about event itself is fine. * how to handle same events * => feature "learning" of same events, double detecting, something like that? * rather pushing event-doers more into own/fav distinction * the devs made workshops in the different groups to teach them how to use radar * hackerspace api could be used for feeding it * establishing trust-ees per language/region or even categories. * groups aren't hierachical. no "metagroups" etc. ### Note about commercial / non commercial 18:04 <xyz> but i wonder how to filter out the obviously commercial stuff that will drop in 18:04 <xyz> as most of these calendars dont filter much 18:04 <abcde> xyz: yeah, i think at the beginning we need to be around it somewhat by hand 18:05 <abcde> i wouldnt drop the commercial ones - but have them also tagged and filter-able 18:05 <xyz> thats prob a lot of work 18:05 <xyz> if they are mixed in the same feeds 18:05 <abcde> xyz: i was hoping - but thats one of the many open questions - that once they are tagged one time, that in the next year the information is already there 18:05 <abcde> and it doesnt need to be done again **18:06 <abcde> anyhow, the definition of commercial is not easy either way...** 18:06 <xyz> yep def not easy 18:06 <xyz> i guess thats something radar.squat.net deals with too though 18:07 <abcde> so i think anyway it needs to be done by hand for the beginning, or at least re-checked by a human 18:07 <abcde> ah, you mean, lots of stuff gets dropped cause its considered "commercial"? 18:08 <abcde> yeah, thats why, i wouldnt like to drop them in total, but keep them somewhat seperate. eg they only show up if you enable them or so. but why dropping, if we "collect" them anyway 18:08 <abcde> also, (and i am still in the process of thinking this through), to really get a worldwide calender, we need to have the commercial stuff somewhat in there, too. 18:09 <abcde> eg in asia, i wouldnt know about any event thats not at least supported by "commercial" entities (except hillhacks and its childs) 18:09 <abcde> but all of them - and they are great events - would go into the "commercial" sorting i am afraid. 18:10 <abcde> life is very different where ppl are still fighting for survival, you know :D 18:22 <xyz> i dont know if stuff is dropped but i know they aim at noncommercial events (the line sometimes can be thin) so maybe theres some experience and suggestions on this issue 19:05 <abcde> xyz: yes, that would be interesting: (1) whats the definition of "commercial" and (2) how to find out and (3) what to do with them then ### categories / filter #### in radar.hackerspaces.org currently: * advice/help/office hours * children * course/workshop * discussion/presentation * exhibition * film * food * meeting * party * Social * work space/diy * sponsored / not-sponsored * profit / non-profit ### We'd need * scraped / not-scraped ### MISC again * so the location thingy is already something we need to change - cause eg "CCC" does various events on various locations. Easterhegg happens at a different location each year :D * note-feature: languages should be rather multiple possible * idea: creating a printed version like stressfaktor has - while entering one could have a checkbox on printed yes/no * also the problem of events / organizers that dont wanna show up can maybe somehow be checked that way? * event http://weboob.org/applications/boobcoming * frontends * [Giggity](https://github.com/Wilm0r/giggity) (on [F-Droid](https://f-droid.org/packages/net.gaast.giggity/)) is a generic Android app for conferences, can accept Pentabarf and vCalendar, or JSON conference description with rooms map, logos… * User Interface * Calendar * [Bootstrap Year Calencar](http://bootstrap-year-calendar.com/#Examples/Full%20example) (limited to year view though) ###### tags: `ccc` `radar` `transmitter`