Talk:Wikimedia Engineering/Maps & Geo Team

From MediaWiki.org
Jump to navigation Jump to search

Focusing discussion here

only productive members ...[edit]

IMO it would be a real benefit if WMF would only take in persons who proved that they deliver "billable contributions", i.e. something donors would click on and give money:

  • one who contributes to wikipedia, wikisource, wikiversity, ... not talk page or meta, real contents
  • one who contributes to openstreetmap, either code, or data, not talk pages, mailing lists.
  • one who contributes code to one of the open technologies used, mediawiki, linux kernel, git, gerrit, mailman, postgres, mysql. not talk pages, mailing lists, bugzillas.

for _all_ roles. we have too many examples recently where it is really painful to interact with persons detached from the community, and anyway getting 5 or 6 digit dollar numbers out of the movement funds donated for free contents. as this is engineering, my suggestion would be to only take open source developers, especially for the community liaison and product management. --ThurnerRupert (talk) 05:51, 5 March 2014 (UTC)

It would be great to have existing OSM community members be part of this team if they are a fit. Do broadcast this to let them know Tfinc (talk) 22:23, 5 March 2014 (UTC)
Every job description inside Tech has had up to three (usually three) sections in the "nice to have/it's a plus" part of it along the lines of 1) experience with open source, free culture, contributions to open source movement; 2) experience with MediaWiki (or a specific open source project that is job related); and 3) experience in the Wikipedia community, working with wikis or other participatory environment.
Is there a reason, if these JD's were open, why that is not enough and those three should be made stricter and elevated to a "must have" category (possibly at the cost of some years of experience as a software engineer)? Or, is this an issue in general problem amongst the current Tech staff? If the latter, with these "too many examples" can you bring them up in detail? After all, I don't know anyone at the WMF who collect 5 or 6 digit dollar numbers (or 1-4 for that matter) for the contributions they make to the wikis in their spare time, "billable" or otherwise, and I know I have to sign a conflict-of-interest agreement every year disclosing even the possibility of such.
(As to the word "billable," I find the implication that there are classes of contributions that are inherently worth a little mercenary and not keeping with the vision of the movement. For better of for worse, I think that if the WMF were motivated to concentrate only on those contributions that net the best ROI in the fundraising campaign, it'd take us down the wrong path—probably one where the only wiki would be english Wikipedia—though in my more cynical moments sometimes I wonder if that isn't already the case :-D)
Take care, --Tychay (talk) 22:58, 6 March 2014 (UTC)

Synergy of OpenSeaMap, Wikipedia, Wikidata, Commons[edit]

Since years there is a aim to build an integrative Water-Sport-Wiki, using OpenSeaMap as geo visualisation and Mediawiki together with content of Wikipedia, Wikidata and Commons. The challenge is to integrate all this branches into a synergetic toolchain and application. Untl now we did not find a volounteer with all tis skills.

What we have:

Wikipedia integration
Port Pilot Book
Diving spots
Canoeing

What we need:

a specialist who can build the tool chain and the interfaces

Then we will get an impressive Watersport-Wiki
and a formidable prototype how to make synergy between the Wikimedia projects :-)

--Markus Bärlocher (talk) 08:23, 5 March 2014 (UTC)

Thanks for a really interesting use case Tfinc (talk) 22:23, 5 March 2014 (UTC)

Great[edit]

This is good news. This is the kind of functionality where maps will benefit everyone. I was already so pleased with the (running) project where a grant was given to work on maps. When you analyse the need for maps within the WMF there are several distinct areas that deserve attention. There is:

  • serving the data from servers
  • internationalising maps
  • link into existing data and maps to overlay them in what will become our standard maps

Maps are used in many places. When Reasonator has a location, either through GPS data or throught the "is in the administrative entity" property, maps are shown. Three maps are shown with increasing detail. There are however many more maps that have their place. Maps of battle fields, historic maps for historic data.

If there is one thing I truly hope for is that the maps functionality will slowly but surely be rolled out. If there is one thing I truly believe it is that perfection will be the enemy of the good. Being able to serve maps is where it starts, supporting and extending the current project is the next most obvious one. Talking with GLAMS is also so very obvious.

In short, thank you for the suggestion to work on maps. I hope it will get its moment under the sun. Thanks, GerardM (talk) 09:21, 5 March 2014 (UTC)

Thanks for your feedback Tfinc (talk) 23:14, 6 March 2014 (UTC)

Very good news[edit]

From the point of view of the Wikimaps project, this will be a needed addition for the development of the Wikimedia environment for the years to come.

The Wikimaps project has set out to enhance the tools for working with historical maps in the Wikimedia family

  • including old maps in Wikimedia Commons
  • georeferencing them to make them available as geographically positioned layers
  • utilizing these maps for extracting geographic features into OpenHistoricalMap
  • connecting the geometries and place names back to Wikidata
  • creating amazing things with this data

We had a kickoff event for a regional initiative within the Wikimaps context - the Wikimaps Nordic - in Helsinki last Friday. We are starting projects within the overall context emphasizing technological development and the regional Nordic context focusing on content. In Wikimaps Nordic there is collaboration with GLAMs: national archives and libraries, land survey organizations, municipalities and cultural organizations in Finland, Sweden, Norway, Denmark and Estonia.

The support for OpenStreetMap is especially welcomed by us, as we are already teaming up with OpenHistoricalMap, mentoring a project in Google Summer of Code together under the umbrella of the Wikimedia Foundation.

We also hope to bridge views about historical place name gazetteers between academic research on the topic and the creation of geographic properties in Wikidata

I am hoping the proposal will be successful, and hope to enjoy the benefits of the infrastructure soon. I would be happy if there would be interaction while developing the resources.

Looking forward to the fun and joy!

This could be a great community partnership. Thank you for sharing. Do you have any office hours or discussion time to find out more? Tfinc (talk) 22:23, 5 March 2014 (UTC)
I'm available at decent hours in EET :) --Susannaanas (talk) 09:34, 7 March 2014 (UTC)

--Susannaanas (talk) 11:47, 5 March 2014 (UTC)

More info[edit]

More focus on user needs and business goals[edit]

Wikimedia is an organization that needs to tightly focus based on resource constraints. Money and energy placed on a new four-person team is time and money not spent on hiring for other positions, and also requires managerial effort to recruit/onboard/guide people.

With that in mind, I think the document needs a much clearer focus on two items:

  1. What exactly would this team be doing?
  2. How does that meet critical user needs?

For the latter, user stories would be helpful. But not just "I'd like to" user stories, but user needs that are critical to our success in engaging readers and editors of Wikimedia projects. What are we blocked from doing that we want to do? Is it build maps in to the Android/iOS apps? Maps on mobile web? Fix geodata for users across platforms to make it performant to use widely and edit?

TL;DR: I think everyone likes the idea of having better maps and geodata. But is it a "must have" in order to accomplish our strategic goals for the next year or two? Let's make that clearer if we think that's true. If it's not true, and this is more of a "wouldn't it be great?" initiative, I'm skeptical that it is worth the resources. Steven Walling (WMF) • talk 13:59, 5 March 2014 (UTC)

When the WMF has a need for money, its track record shows that it can find all the money it needs. So the question is more about energy and ability to manage. When you consider other positions, the question is very much about how do you measure the value of the work done in comparison to other things that are ongoing. There are no objective criteria and several projects just happened.
The first thing that would be done is fulfil promises made in the past. Maps and at that OSM maps were promised to be hosted within the WMF domain. This never happened.
The second thing to consider is that maps transcend Wikipedia. It has the potential to help in many activities that we are known for. Projects like "Wiki loves Monuments". There are projects to ask people for pictures that are near to where they are. Good maps help. Good maps explain many things much better than a narrative. We have examples of maps of battle field over time. We have maps of the development of cities, the growth / waning of empires.
The list is endless. The good news is that it is language agnostic. Maps will help any project in any language. Maps can be combined with data from Wikidata, images from Commons.. Maps may be liberating us from the Wikipedia monomania and English at that. GerardM (talk) 11:42, 6 March 2014 (UTC)
While I also like the synergies and collaboration that will come out from proper maps/geo support very much and I believe that building a Free maps infrastructure will benefit the movement greatly, I have to strongly agree with Steven here. The list of items to do are fairly generic, almost nebulous. Where are the "4 engineers, 1 ops" figures coming from? Four/Five/Seven FTEs is a lot of manpower in our size; most teams/subteams are far less than that, including projects that are far larger and more complex.
I also don't see the rationale for implementing these functions in a separate team, at least immediately. Do we have existing engineers/PMs tasked with the infrastructure & product that have expressed that it's impossible to reach their goals with the resources they have in their existing functions?
Finally, I believe that most of the frustration arises from the lack of a Maps tileserver infrastructure, despite long-time promises to the contrary. This is entirely in the Operations department's charter and it should remain there, IMHO. This has not happened for various reasons that have to do with resource constraints, prioritization and bad management (and yes, we engineers share the blame for that) but I think this has finally started to take off recently and I don't think just throwing people into the problem will solve this immediate issue (mythical man-month). Faidon Liambotis (WMF) (talk) 19:48, 6 March 2014 (UTC)
Thanks for the feedback Faidon. A lot of this proposal goes back to my discussions with Brandon Black and him articulating that the only way that the OPS team could support this is by having dedicated staff to build out the infrastructure. Have you had the time to talk to him about this? It's always murky to me to know how well this information is passed through the ops team. And as for the product groups Mobile Apps, Mobile Web, and Multimedia have chosen not to take on projects or have had to adjust them as a result of not having any maps infrastructure. The web team has had to take up its own cycles to build GeoData. On the community city we'd had dozens of discussions and requests from the mapping and geo community and for the most part we've scrounged for resources, delayed, and generally haven't delivered well at all due to lack of resources. Tfinc (talk) 23:12, 6 March 2014 (UTC)

Copyright compatibility[edit]

One of the ongoing problems between Wikipedia and OpenStreetMap is copyright compatibility. OSM uses a database license, Wikipedia does not. There are good technical reasons for that, but there they are. Also, Wikipedia has encouraged people to pull lat/lon information from Google aerial imagery. Some OSM users make the claim that that infringes Google's Terms of Service. It would be kinda awesome if Wikimedia could talk nicely, and on a high (executive) level, to Google and get an official statement that using Wikipedia editors using Google aerial imagery to determine a lat/lon point is not against Google's Terms of Use.

I agree with Steven's point above. What is the value-add? Is the goal for the WMF to serve up its own map tiles? For all comers or only for its own mobile client? Are the tiles rendered generically or with a special marker indicating the location of a Wikipedia entry? Some Wikipedia pages describe linear features which are already laid-out in OSM. Is a goal of this project to create greater cooperation between WP editing and OSM data? Specifically, will this map project create a way for a WP editor to make reference to an OSM way or relation and have that automatically be turned into a cross-reference? --RussNelson (talk) 15:51, 5 March 2014 (UTC)

I'm not aware of any copyright problems between Wikipedia and OSM. Could you elaborate? Also, why would we need to create a cross-reference if we could simply embed the maps within the Wikipedia articles? Wouldn't that be a better solution? Kaldari (talk) 18:42, 6 March 2014 (UTC)
Anything that involves mixing Wikipedia and OSM content can run into copyright and database rights issues and should be done with those issues in mind. Some background:
Aude (talk) 09:08, 7 March 2014 (UTC)

Yes, "an official statement that using Wikipedia editors using Google aerial imagery to determine a lat/lon point is not against Google's Terms of Use" would be great! A lot of WP coordinates are from Google - and OSM has a rule to not use this. So if we get a such permission, this WP coordinates can be used in a synergetic way. --Markus Bärlocher (talk) 14:29, 25 March 2014 (UTC)

Go Big or Go Home[edit]

You know what we need for a real breakthrough? A technology that doesn't exist, but which we're >< close to being able to do. It's one thing to look at photos and maps and prose. But all of this exists in the real world. You can GO to these places and see things.

What's really needed is a way to georeference the pictures, maps, and prose, so that if you are out and about, and reading this document on a tablet with a GPS and compass, you get an indication of where you are. Everything that is georeferenced would have a little red arrow pointing at you (if you're outside the view of the photo or map), or pointing in the direction you're holding the tablet (if you are within the view of the photo or map). Or if it's a prose description of something, there would be an arrow pointing in the direction of the thing described.

Three things are needed that don't exist: 1) A document reader that will let you zoom in and scroll around on images (pictures and maps) whilst displaying your location and view, 2) a standard for how to write documents that the reader knows how to interpret, 3) content written to that standard, and 4) tools to help people write that content.

Can you imagine how AWESOME that would be?? Look at a picture of a building, and see which direction it's in. "Oh! This says it's right behind us!"

IMHO, that's what this team should be working on. Go big, kick ass, or go home. --RussNelson (talk) 19:15, 5 March 2014 (UTC)

...and a statement[edit]

I would like to say a word or two about the purpose of the Wikimaps project and why it is interesting to work within Wikimedia.

Wikimedia has a promise of being a maker space of digital knowledge. A place for collaborative enquiry, creating the sum of all human knowledge. Location is a key component of knowledge. Why would you decide not to support it?

We do GLAM work to persuade memory organizations to invest in Wikimedia, to offer their materials for this space for people to engage with. We ask people to invest their time and expertise. Engament follows opportunities and the promise of what Wikimedia has to offer the people. The enthusiasm of people working to rediscover their cultural heritage should not be left to the commercial actors. What is the goal if not enabling this?

We are working to collect use cases for the Wikimaps environment and will report back.

--Susannaanas (talk) 08:41, 6 March 2014 (UTC)

A specific proposal[edit]

Per the "More focus on user needs and business goals" discussion above, I would like to make a specific proposal. I would like to propose that MediaWiki implement a "Map" namespace. The actual content of the namespace would be JSON data specifying:

  • Either a tileset or local image file
  • A coordinate bounding box
  • Data for points (places) that should be highlighted

The JSON data would not be edited directly, but through some sort of custom interface (similar to Wikidata) that allows interacting with the map. Once the map is defined, it could then be embedded in an article/page just like an image or video. Kaldari (talk) 19:14, 6 March 2014 (UTC)

I think the local image part is crucial to getting wider community support.
We should check what the OSM community has in terms of formats and requirement on formats, that might help us to achieve this. What about we ask that company (mapbox ?) that made their new online iD map editor, see if they have any advice for our efforts ? And I'm sure wikimaps have some ideas, with their efforts regarding overlapping with historical maps etc. TheDJ (talk) 21:21, 6 March 2014 (UTC)
Relevant templates that are currently responsible to do similar stuff: en:Template:Location_map, en:Category:Geobox2 and a few more in en:Location_map_templates. TheDJ (talk) 21:55, 6 March 2014 (UTC)
There are efforts that the Wikimaps project is tackling with, that relate to the mentioned:
  • Creating a Template:Map, that would store the metadata held by the archival institution and allow rough georeferencing of the image much like in this piece of software or more precisely with MapWarper that has an experimental instance at the Labs. We have some preliminary specifications for the template. We have also discussed about including the template in the GWToolset.
  • Integrating working with the maps in Commons to the iD editor in a seamless user experience. Some concept sketches here.
It would be great to initiate workshops around these issues at the Zürich hackathon. --Susannaanas (talk) 08:21, 7 March 2014 (UTC)
We should probably also see if we can make the infra in a way that it can be reused for Annotations for images feature that has long been desired. TheDJ (talk) 11:37, 7 March 2014 (UTC)

Scenarios[edit]

Re: "Update WikiMiniAtlas to be supported by internal tile service, Currently maps are slow and don't even load consistently"

  • If we had a fast internal tile service and labs osm database, I am quite sure the developer of WikiMiniAtlas could remedy this. Right now things are constrained by the resources (or lack) provided.

Re: "Update synthetic maps like File:California_county_map to be easily creatable"

(and see you are are aware of WikiMaps)

Where there is work being done in the community, I think the resources of WMF engineering are better focused on providing the infrastructure that would allow the community to make their tools better. For features, focus on things not being done by the community like geo stuff in mobile. Aude (talk) 09:02, 7 March 2014 (UTC)

I agree in principle, but where it comes to INTEGRATION, staff is key. And much of the efforts of volunteers have been stuck with their progress at the point of further integration. S o that is where staff definitely can help and be valuable. TheDJ (talk) 09:25, 7 March 2014 (UTC)

My support[edit]

I strongly support this proposal. WMF needs a team inside the organisation that understand everything about maps from the visions for new features, data connection, UI of maps, existing software solutions and at the end efficient server operation. If you ask why Wikipedia should invest in maps, you should ask yourself why google invest billions into maps. Geographical knowledge is a fundamental part of knowledge, to understand the world. Maps are also an important part of multimedia inside a modern website.

The way of low effort in the last 1-2 years doesn't work. Developers from the community were blocked by to many misunderstandings and broken promises. If I look in the tiles delivering statistics of Toolserver the tile deliveries decrease by 50%, but in the same time OSM grow and grow. Also in absolute numbers the map usage inside Wikipedia could be much higher. We are in pro-mile area that visitors use our maps inside a geo-coded article, I see a potential of 5-10%. So Wikipedia is in this geographical area far away from it's potential.

We should also work to make our data (Wikipedia-coordinates, geocoded Commons-images, geodata inside Wikidata, ...) easily available for external developers.

In the past google maps had a Wikipedia-Layer, they removed it (IMO to concentrate more on objects of commercial interest like shops and restaurant with that they want to make money.) So I think there should be still a big demand for maps with cultural content that nobody could better fill than Wikipedia.

The development in maps technology is very fast and the specific wikipedia-topics (Merging OSM and Wikidata, Multilingual Maps, Vectorrendering for individual maps, touristic informations, mobil...) are very complex. So only to have 1-2 operators is not enough.

Last word: It will be important that the new team will not try to reinvent wheel, but cooperate with long time, volunteer developers. Defining together projects for new features that are to complex for volunteers and on the other side support them. --Kolossos 23:48, 8 March 2014 (UTC)