Hi all, first of all I regret very much that I won't be able to make it to the Zurich Hackathon, but Idaho, USA is a bit too far and I have just switched to a new job. Anyhow, I'd just like to chime in and support what Derk-Jan said above. Even though the WikiMiniAtlas, a dynamic map, is my project I think it is very important to recognize the way maps are used in Wikipedia by our contributors. While this crowd is probably rather interested in the technical side of dynamic maps, mapping in Wikipedia is very much an 'editorial' process. Dynamic maps are nice to provide a quick geographical context and show nearby points of interest, but I seriously doubt they will replace the types of maps we have in articles today (with a few exceptions maybe). We have political maps on many administrative levels, we have maps of biomes and species distributions, geographical and geological maps (e.g. watershed maps), climate maps, historical maps of battles and population migrations, animated maps showing the evolution of empires, extraterrestrial maps, and so on and so forth. Don't get me wrong OSM data is awesome, and very rich, and it is a fascinating and fun work to integrate as much of it into Wikipedia as possible. But we have to keep in mind that OSM offers only a limited scope in the vast array of map applications. And while simplification and automation sounds appealing to us, what the mapping community will see is most likely a loss of editorial power compared to custom maps. Best, Daniel
On Fri, Mar 7, 2014 at 5:36 PM, Derk-Jan Hartman d.j.hartman+wmf_ml@gmail.com wrote:
On 8 mrt. 2014, at 00:47, Russell Nelson russnelson@gmail.com wrote:
Maps great, yes. Here's what *I* think Wikipedia needs. First and foremost, OSM is a database, just like Wikipedia articles are data (don't get me started). It's a foreign database, which means that you have the foreign key problem. Wikipedia articles all have a published unique name. Perfect for a foreign key. Articles can get moved, but in the ordinary course of editing, they don't. In OSM, the key "wikipedia" is used to make reference to a Wikipedia article. It can be the full URL, or simply the article name (latter is recommended). http://wiki.openstreetmap.org/wiki/Key:wikipedia for details.
I agree, somewhat, but I think this is not 'Wikipedia' per se. It's as much a need for OSM, as it is a need for Wikipedia. And articles do get moved (as much as OSM nodes with tags on them get deleted).
Linking to OpenStreetMap entities, on the other hand, is somewhat more abstract. OSM entities (nodes, ways, and relations, all fruitful for Wikipedia linking) have a database ID. The trouble that the ID is only as persistent as the database entry, which can change through as simple a change as adding a bridge to a road (one database entry becomes three -- the original, a new short one which is the bridge, and the remainder of the road, as a new way). Probably the most persistent way to link to an entity is to give a lat/lon of a node, and give the name of the node. That is enough information to recover from any editing (if the node gets moved or deleted, there be congruent nodes nearby, and if the name is changed, you still have the node's location). Particularly if the entity has a Wikipedia article name; a bot can audit both databases against each other.
Wikipedia currently has only points in articles. At most you can say "The subject of this article is at this point." which is insufficient if the point is an area (in OSM, a self-terminating way), a way (e.g. road, railroad, river), or a relation (a collection of related ways, e.g. a lake and its islands, or a bus route). The Wikiuniverse could reinvent OSM's system for identifying places, or it could bite the bullet, lean more heavily on OSM's data, and make article references into explicit OSM references. Note that these references would not be OSM specific -- any map is going to have lat/lon and names of entities. It's just that the Wikipedia editor that helps make these links would use OSM lat/lons and names.
Well first of all, we should get away from the concept of 'article'. It's really clear now that most of this data will soon live in WikiData. Coordinate values have already been added, and it's only time until most of those existing values have been moved into WikiData and there is the OSM relation ID property to link from WikiData to OSM. For entities/concepts WikiData is the only thing that will be relevant real soon. And a bot can easily detect when those break. So what you want in this regard is already in process.
Then, once Wikipedia articles can make reference to OSM objects, and you have a OSM-derived Wikimedia-hosted set of map tiles, then including a dynamic map in an article is simply a matter of a single command. And an article which is a collection of articles can make a map with references to those articles, again, with a single command.
This is thinking way too simple. Would it be nice to be able to do this ? Sure, but we already have dynamic maps. WikiMiniAtlas, the german OSM map. Those even have area highlights these days. Can they be better ? sure, is that something difficult, is it really needed ? No not really x2. Wikipedians are much more focused on making static maps usually. Or rather, they often want very specific types of maps, resulting in them using custom designed static maps. These custom maps have different base layer types for different types of articles (and sometimes for different cultures cross wiki), and very simple highlights in them (less is more). Or they want to mark the spot of an event (planecrash/hurricane path etc). If we do not recognize the target group, then the integration of the feature will fail. Also, there are many problems with dynamic maps in articles (print, mobile, loadtime, fallback for users without JS). Can we do many cool things with maps ? sure. but where is the most valuable place to start ? That is a much better question I think.
Personally, I think the power is in annotation. The success of FA, Commons, WikiData and OSM have shown that our users love to not just 'collect' data, but to do the curation, annotation and restoration of this data. Doing annotation using a simple OSM renderer as our starting point could be step 1. We need support for 'static' image maps to generate the buy in of people to start using this feature (step 2). With enough buy in, we can start going for the long haul, which is to build a very complex map renderer with a lot of different baselayer configurations, that might help us at some time achieve to phase out the static maps. But saying we need 'dynamic maps' is way too narrow and technical view.
The other thing we could build, would be the 'foursquare' microcontributions to our mapping data (can you pinpoint this location on this OSM map, there are five locations in this area, which one best represents this article) and then with enough confirmed information set/add/correct/suggest the OSM value in wikidata again. Things like that (gamification). I think that is where we have a lot of potential.
DJ
And .... here's the money shot: your location gets shown on the map.
BTW, Special:Nearby is broken. Android phone, Chrome Browser, GPS on and locked (Ingress is happy), "Quiet out here ... there weren't any pages found nearby." Yet I could throw a stone and hit the location of a page AND my Google Glass app is happy to report that a page is very close to me.
On Fri, Mar 7, 2014 at 4:55 PM, Arthur Richards arichards@wikimedia.org wrote: Somehow wikitech-l got dropped form the recipient list of this thread; I know some of the OSM folks are subscribed. Anyway, re-added wikitech-l to this thread.
On Fri, Mar 7, 2014 at 1:53 PM, Jon Robson jdlrobson@gmail.com wrote: Sounds like a good idea. If there is no objections to my potentially crazy idea I will drop them a note on their mailing list...
On 7 Mar 2014 13:27, "Arthur Richards" arichards@wikimedia.org wrote: Dunno if any of the OSM-y folks are planning to attend but I bet this would be up their alley. At the very least, it would probably be good to get their input on a project like this.
On Fri, Mar 7, 2014 at 11:44 AM, Jon Robson jdlrobson@gmail.com wrote: Dan awesome! Glad there is some interest - this should be a lot of fun! :-)
On Wed, Mar 5, 2014 at 1:20 PM, Quim Gil qgil@wikimedia.org wrote:
fyi
-------- Original Message -------- Subject: [WikimediaMobile] Zurich Hackathon: Creating a map namespace Date: Wed, 5 Mar 2014 12:54:52 -0800 From: Jon Robson jdlrobson@gmail.com To: Wikimedia developers wikitech-l@lists.wikimedia.org, mobile-l mobile-l@lists.wikimedia.org
This may be extremely ambitious, but I'm keen to kick off development around the creation of a map namespace during the Zurich hackathon.
The goal would be to setup an editable map namespace that could be used for a variety of things, one of which would be adding a map view to the Special:Nearby page provided via the mobile site. The goal is a proof of concept not necessarily anything production ready (but that would be great if we could get to that point!)
Please let me know if you would also be interested on hacking such a thing - https://www.mediawiki.org/wiki/Z%C3%BCrich_Hackathon_2014/Geo_Namespace
- or if doing so would be a terrible idea (but if you have to go down
that route please provide constructive reasoning on what would be a less terrible idea)
Excited to hack on cool things in Zurich!
Mobile-l mailing list Mobile-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/mobile-l
-- Jon Robson
Mobile-l mailing list Mobile-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/mobile-l
-- Arthur Richards Software Engineer, Mobile [[User:Awjrichards]] IRC: awjr +1-415-839-6885 x6687
-- Arthur Richards Software Engineer, Mobile [[User:Awjrichards]] IRC: awjr +1-415-839-6885 x6687
Mobile-l mailing list Mobile-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/mobile-l
Mobile-l mailing list Mobile-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/mobile-l
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l