Accueil / Blog / Métier / 2013 / Feedback on the State of the Map 2013

Feedback on the State of the Map 2013

Par Frédéric Bonifas — publié 26/09/2013
Notes on the OpenStreetMap's annual conference, to which Makina Corpus was participating.

This year I was at the State of the Map, the yearly OpenStreetMap (OSM) contributors meeting, in Birmingham. The sessions were once again of very high quality and the meetups especially interesting. In this blog post, I present the talks that mostly drew my attention.

Photo de groupe

Picture by Chris Fleming

Contributing to OSM

  • Guy Collins shared his thoughts on the complexity to define when a map area is completely mapped out. The map is sliced by Voronoi polygons created around the postboxes. He also created a model of completeness in several steps, taking into account the different objects that can be mapped, like roads, landuse or postal addresses.
  • Several talks dealt with the gamification of OpenStreetMap contributions: some applications are already online but scarcely used. While games aiming to contribute to OSM could attract new contributors and improve the quality of existing data, there are many challenges: how to prevent cheating? How to avoid high-volume but low-quality contributions? How to avoid discouraging experienced mappers that contribute on technical subjects and not on adding lots of new objects?

Using OSM data

  • Interesting thoughts of the AND company, that donated data to OpenStreetMap and that is also using data from the project, on the share alike clause of the current license. This clause is often hard to observe for a company using OSM data: what to share back? To which point share the modifications that have been done? Which format to use? Clearer rules from the community on this clause could allow to receive more information back from the commercial users of the OSM data.
  • Presentation of Open Historical Map and the challenges emerging while building a spatial and temporal database, re-using several components of the OpenStreetMap infrastructure.
  • Presentation of Photon, a geocoder using OpenStreetMap data and based on Apache Solr. Some questions raised the advantages and drawbacks of using ElasticSearch instead of Solr, without answers for now.
  • Presentation of Tileman, a cache and tile server working with Nginx.
  • An introduction on how to build its own vector tiles by Norbert Renner. The technologies used are:
  • An example of application developed with these tools was shown.
  • I was also able to introduce Moodwalkr, the pedestrian routing application I have been working on in the last months.


  • Several talks on the place of women in the OpenStreetMap community, the low number of them actually, and the means to address it. In particular, have a look at this presentation and this other one.
  • Frederik Ramm presented an interesting analysis of the evolution of the OpenStreetMap project through its mailing-lists and the other means of communication available: the wiki, the forums and the help website. He remarks that each mailing-list - 132 of them exist in different languages or on different topics of interest - has created its own community, talking sometimes on the same subjects but at different times.
  • humoristic video, Gary Gale not being able to attend the State of the Map, on the 8 steps that lead its map of rude place names to go viral on the social networks.

Future of OpenStreetMap

  • The website has already evolved a lot in the past months, especially thanks to the work of some MapBox employees: new iD editor, redesign of the tools related to the map, new user menu et new registration page. Several other improvements are expected: revamped user profiles, creation of user groups and global redesign of the home page. To follow this, you can look at this Github repo.
  • A new version of the API, that would be the 0.7, has been in discussion since even before the launch of the API 0.6, 4 years ago. However, the process to define the new features is stuck because of queries in high number, too complex and often too specific. Matt Amos therefore proposes to relaunch the discussions on a new version of the API with lower expectations. The two main features could be a better management when uploading changesets by implementing a queue, and the creation of a new datatype for the polygons. Jochen Topf also spoke on this topic of a new datatype for polygons, that would solve the following problems:
    • ambiguity on what is or not a polygon currently, solved only by looking at the attributes of the object;
    • several ways to model a polygon, by a polyline or a relation;
    • the management of coastlines and rivers is currently not satisfying.


See you next year !

Voir aussi
Web mapping : comparaison des serveurs de tuiles vectorielles depuis Postgres / PostGIS Web mapping : comparaison des serveurs de tuiles vectorielles depuis Postgres / PostGIS 30/07/2020

Un ensemble de serveurs de tuiles vectorielles basés sur la fonction ST_AsMVT() de PostGIS sont ...

SIG : Mettre en place des tuiles vectorielles SIG : Mettre en place des tuiles vectorielles 07/04/2020

Pour les données images « rasters » comme pour les données vectorielles, des techniques de ...

Une Cagnotte reversée à la fondation OpenStreetMap a été organisée par Makina Corpus lors du Paris Open Source Summit ! Une Cagnotte reversée à la fondation OpenStreetMap a été organisée par Makina Corpus lors du Paris Open Source Summit ! 28/01/2020

Les 10 et 11 décembre derniers, Makina Corpus était présente à la 5ème édition du Paris Open ...

SIG : Préparation de données pour la création de tuiles vectorielles 30/01/2020

Pour servir des données sous forme des tuiles vectorielles une préparation est nécessaire comme ...

Recherche et développement d’indicateurs pour l’aménagement d’un territoire Recherche et développement d’indicateurs pour l’aménagement d’un territoire 14/01/2020

Nous explorons l’intérêt de calculer automatiquement des indicateurs d’aménagement du ...