Venue maps in Kongress
With Akademy 2024 hosted in a venue with OSM indoor mapping, what happens if we put KDE’s conference companion app Kongress and Itinerary’s train station indoor map view together?
Venue maps
Out of the box we get a multi-floor map, which is already an improvement compared to just showing building outlines on a conventional (outdoor) map. Visually it’s still far off from handmade conference venue maps though.
This is using vanilla OSM data (via KDE’s raw data tile server), which is following upstream within 24 hours. Any work done on the map data is thus not only benefiting a specific event but all OSM users.
Finding rooms
Using OSM raw data and client-side rendering rather than pre-rendered raster graphics allows for additional ways to interact with the map or browse the data.
Itinerary has a nearby amenity search feature for example. What you are also often looking for at a conference venue are rooms though. We can apply the same concept here, with a few adjustments on which data to display and how to group it.
This is also the foundation for matching room information from the schedule to the map. That’s not implemented yet in the app though, as we’ll need the actual schedule for testing that.
Custom styles
For a conference in e.g. a university building only a small subset of rooms might be relevant for an event. Custom made venue maps often highlight those somehow, e.g. with different colors.
As our indoor map renderer uses MapCSS stylesheets to control the visual appearance, that’s easy enough to do here as well. And since MapCSS stylesheets can include other files, we also only need a small event-specific addition that then includes the default style.
Custom content
While stylesheet can do a lot, they can’t generate new content on the map. That’s also needed though, as there might be structures that are only there for the event and thus too short-lived to be added to OSM.
This can be solved by applying an OSM changeset file on top of the loaded map data. That’s basically a “diff” for the map data with event-specific additions, changes or removals.
This way event-specific content becomes part of the map data itself rather than being in an app-specific overlay. We can therefore make use of all the existing infrastructure for the custom content as well, from support for multi-lingual labels to being searchable.
Outlook
And there’s still more we could do.
Time-dependent elements
Some elements might only be available or applicable for a certain time of the event, or certain rooms might have different purposes on different days.
OSM has a complex mechanism to model time-dependent information, using opening hours expressions. We can parse and interpret those as well, but their use is currently limited to showing actual opening hours for amenities or shops. This is accessible to styles, so we can e.g. gray out currently closed things.
Generalizing this and also allowing to interpret opening hours expression on arbitrary tags in combination with custom styles should not be too far out and give us a really powerful mechanism to model time-dependent features.
Routing
Kongress also has our experimental indoor router integrated. For the Akademy venue this works ok-ish (apart from access to the elevator, the map data quality around that isn’t good enough yet).
An important still missing part here however is the ability to select and customize routing profiles, as routing becomes particularly useful when considering mobility constraints.
Also note that we are only talking about routing here, full navigation is much further away still as we need to have working indoor localization for that as well, which is a much harder problem to solve.
Do we actually need all this?
For a comparatively small event like Akademy in a relatively simple two-floor venue, probably not. But think about your first visit at a massive event like FOSDEM, or consider stairs being an insurmountable obstacle. Having support with finding your way around becomes much more important then.
Akademy is a great opportunity to experiment with this though, both from the software and the mapping perspective, and to show what’s possible with the data and technology we already have.