With three days of three parallel conference tracks with tightly packed 20 minute slots I only got to see a small subset of the talks, focusing on the topics most relevant for KDE Itinerary. Some takeaways for me:
- OSM core data model evolution: The initial steps discussed here aren’t directly impacting KDE’s uses of OSM data yet, the possible improvements for more efficient and accurate tile expiry are something potentially interesting for our raw data tile server though (although we currently don’t implement any form of tile expire yet).
- Indoor positioning (with GPS usually not available inside buildings there is no similarly prevalent solution yet, let alone one that works without needing extra infrastructure in the building and without requiring non-standard/not-yet-standard hardware in phones): Two possible approaches for this were presented, one using common inertial sensors and map matching to compensate the drift, the other using a camera and a SLAM-like algorithm. Nothing published yet unfortunately, so we have to see how well those actually perform in practice.
- Indoor map data: I found it particularly interesting to see for which very different usecases people need the same kind of data, from navigating through a train station to city-scale earthquake risk assessments with a scary level of detail and accuracy.
I also did my first major conference talk in German there, which I hopefully managed to do without using an English term for every other word.
Public institutions and FOSS
There were plenty of presentations from public administrations and universities, for all of whom it’s no longer a question whether or not to publish their data and source code under an open license. There is a bit of a selection bias by the event here of course, but given how often the Public money? Public code! slogan of the FSFE was mentioned, it would seem campaigns in that direction had an effect. And that’s absolutely great.
It doesn’t stop there though, I think we also can help those institutions with how to do this properly. A one-time code dump on Github once a project is finished isn’t the ideal way for example.
- Start in the open, and do as much as possible in the open. Not just code, this includes communication, requirements and goals, design and planning, etc.
- Think about how to deal with input, feedback and contributions from the outside. This needs time and effort to review and process. It is also not unlikely that outside parties have slightly different goals, priorities and timelines, ie. you need to have a plan for project governance.
- Consider how things can live on after your funding/grant or research project runs out. Who owns the repository and other project assets? Who continues server operations? What happens to app store accounts or signing keys?
This might seem obvious for people active in larger communities like KDE or OSM, but that’s in part because we are used to having a long-living umbrella organization handle a lot of these issues for us. Umbrella organizations could be part of the solution here as well, either any of the existing ones or something new e.g. on a national or EU level for public administrations.
It’s also good to see that some institutions are indeed considering those aspects and see FOSS as more than just a grant completion form checkbox.
BoFs and hallway track
There were also plenty of interesting discussions, from BoF sessions to chats in the hallway, on a wide range of topics:
- Getting the various people looking at indoor map data connected and working together on the open questions around data modelling.
- Ongoing work and new findings around indoor positioning and routing, as well as legal questions around obtaining indoor map data.
- Modelling and collecting accessibility-related attributes. From the KDE perspective accessibility usually means building accessible software, here it’s about software assisting with navigating the more or less accessible real world.
- Tools for aligning OSM data with other data sources of varying quality, e.g. the official data sets of public transport stops and their respective identifiers. That’s relevant for matching public transport journey data to the exact stop position.
- Handling of transient map data, such as temporary setups around conferences or other events. Imagine we’d use our indoor map component in Kongress and we want to see things like a temporary added registration desk there and be able to show event-specific room labels (e.g. “BoF room 1” instead of “Room 1.234”), on top of the OSM base map of the building.
- Various approaches for moving beyond raster tiles. The raw data tile setup we have at KDE for this is a bit of an unusual solution compared to what else is out there, but something in that direction could be one piece of that puzzle. Either way, being involved in upstream discussions on that topic makes sense, the more we can align KDE’s infrastructure with what OSM does the better I think.
- Putting things on a map (quite literally here) makes things visible, and that has real-world implications. This can be good (visibility for a issue you care about) or bad (security implications for e.g. certain vulnerable groups), and that assessment might even depend on who you ask. How should decisions be made in that context, and how to deal with the responsibility this entails?
Another recurring topic at any FOSS event is sustainable funding. Project-based low-threshold grants are available from a number of entities (NLnet, Prototype Fund, mFund, etc), but their short term nature make those only viable for a very limited subset of the people you’d ideally want to see funded for working on FOSS projects and infrastructure.
The deeper I dive into the OSM community the more déjà-vus I get, and it seems there is probably more we can do in terms of inter-community collaboration/knowledge sharing in the FOSS/open data world, in particular around adjacent/supporting topics, ie. things we all have to deal with in order to achieve the things we actually want to do.
Simple example: we shouldn’t all have to learn the hard way that Jambra conference mics, BBB and your average university seminar room don’t really get along very well out of the box.
Event management, hybrid events, funding sources/grant programs, getting useful results out of GSoC, PR and lobbying on various levels, legal and licensing, infrastructure operations, the list of potential topics is long. For some aspects there have dedicated entities like the FSFE, but for many others this seems to be much more ad hoc.
Not that I have a good idea on how to address this, but it does seem like something where everyone could benefit from more exchange.
Either way, attending events outside of our own bubbles certainly helps with that :)