Two weeks ago i participated in the OpenStreetMap Hack Weekend in Karlsruhe. I was not able to take part in many OpenStreetMap related in-person meetings in recent years due to a mixture of lack of budget and scheduling conflicts. So it was nice to meet various people there i had not seen in years. I also managed to get some actual work completed during the weekend, which i will introduce in the following. Historically, i have rarely managed to do so at hack weekends because i have mostly been working on map design – where work is frequently difficult to fit into a limited time frame. I also made progress on some map design tasks this time. These are going to require some more work though and likely will be presented at a later date.

What i managed to complete at the Hack Weekend is generating a taginfo.json file from Styleinfo data and this way get OpenStreetMap Carto into Taginfo projects. This had been on my todo list since creating Styleinfo in 2022 but so far i had never gotten around implementing that.

OSM-Carto in Taginfo

Styleinfo analyzes map styles without much a-priori information on the map style in question. It does so by systematically rendering samples of features with different tag combinations at different zoom levels and checking what the map style actually displays then. I explained this in more detail back then in a blog post. This is expensive to do, analyzing a complex map style like this can take days and it is not a hundred percent complete – some tags and tag combinations tend to be missed by the heuristics. But it produces very detailed data on how the map style actually renders things.

Taginfo includes information of data users (like map styles and various software) on how they interpret OpenStreetMap data based on structured information on the data use provided by the project through taginfo.json files.

Apart from general information on the project, that JSON file mainly contains a large array listing the tags the project interprets with optional data on how that tag is interpreted in the form of a description text, a small illustration image or a link to further information (or all of these). Originally, i had assumed that this allows for exactly one array entry per tag + feature type combination (feature type being node, linear way, polygon or relation). This would have been somewhat difficult for secondary tags like access=*, which a map style tends to interpret very differently in different contexts.

Interestingly though, Taginfo accepts an arbitrary number of entries for the same tag, documenting different interpretations of that tag. At this time i used that possibility to provide information not only on the >600 individual tags OSM-Carto interprets, but on each of the >2500 tag+feature type combinations. I could have further differentiated out the zoom levels (which would have meant ~15k different rendering variants) – but decided against it. If you want this detailed information on the map style, you should use Styleinfo.

The other interesting thing about how Taginfo deals with the projects is that it does not cache the small images illustrating the tag use, but includes them directly from the URLs specified in the taginfo.json file. That was fairly unexpected, considering nothing about this is mentioned in the privacy policy – the OSMF has some homework to do here. So, when you browse information on projects in Taginfo be aware that the images shown in the table there are directly requested by your web browser from wherever the project provides them from.

The bottom line for you as a Taginfo user is that you now have the information on what tags are interpreted by OSM-Carto, and how they are interpreted, available in Taginfo – including small rendering thumbnails and links to Styleinfo. I also provide links to the taginfo.json files for the other styles featured in Styleinfo, but i have not submitted them for integration in Taginfo. This is up to the maintainers of these projects. For my own alternative colors style i plan to do that, but first i want to update the analysis in Styleinfo since a lot of things have changed since 2022 – much more so than in OSM-Carto.

Why so late?

Why did it take almost three years to make this available? As mentioned, this had been on my todo list since i created Styleinfo. But my free time is limited and this did not have high priority. I developed Styleinfo for my personal education and as a strategic investment into systematic testing in map design. I made the results of this available for everyone else to study and use as reference to the map styles analyzed. But i had fairly little personal interest in also making this information available in Taginfo – hence the low priority to do this in my free time.

But don’t make the mistake to conclude now that nothing could have been done to expedite this. Because i was talking only about what i do in my free time. Paid time is a different matter. It ultimately took me about one work day to implement this. If you had asked me for an offer to do this in 2022/2023 i would probably have calculated a bit more (and to be accurate: I had already read up on the taginfo.json specs before the Hack Weekend). The bottom line is: You could have gotten this already in 2022/2023 simply by moving it from my limited free time todo list to my paid time schedule – with a rather minimal budget.

Why am i pointing this out here? Because i think it well illustrates a larger issue the OSM-Community currently struggles with. It is beyond doubt that money and economic aspects play an increasing role in OpenStreetMap. You can criticize that because you wish back the good old days when OSM was fully under control of hobbyists. Alternatively, you can dismiss that criticism as naive and backwards. But both of these perspectives miss some really important aspects.

Money is neither inherently good or bad for OpenStreetMap, it depends on how you use it. And one of the things people in positions with influence on budgeting decisions (both in private businesses and in organizations like the OSMF and FOSSGIS) seem to almost universally fail to see is that a huge part of the human potential (available time and competencies) in the OSM community exists in the form of the following two groups of people:

1. Pure hobbyists who invest their free time into the project, but who exclusively do this out of intrinsic motivation. It is almost impossible to positively motivate those with money but very easy to de-motivate them with it. Some of these will be open to being offered economic benefits (like travel cost coverage, paid assistance), but very few will be willing to ask for such. Attempts to try to influence their decisions on how they invest their free time with money – either directly or indirectly – will usually have a highly negative effect on motivation.
2. People who have some level of professional connection to OpenStreetMap, but who are also investing their free time into the project separately from their professional time and who – like the pure hobbyists – spend their free time in the project purely out of intrinsic motivation. For these people, time spend on OpenStreetMap can either be professional time or personal time and they separate between the two. In their decisions on how to make use of their free time, these people tend to act like the pure hobbyists, while on their professional time they adjust to market demands.

For symmetry you can define two other groups of OSM community members:

3. People who – like the second group – spend both paid and unpaid time on the project, but who do not really separate between the two. These often come from one of two backgrounds: (1) Idealist professionals who took a job related to OSM out of idealism and who, therefore, get involved into their work beyond the time they are actually paid for and (2) Former pure hobbyists who have turned their hobby into a profession. In contrast to the second group, this group is open to adjust also their free time involvement to economic forces and tends to be much more willing to proactively ask and lobby for funding for their free time work than the second group.
4. Pure professionals who are involved in OpenStreetMap in their paid work, but who are not involved beyond that.

You can easily see that these four groups are not inherently discrete, they form a continuous spectrum. And, as you can probably imagine, i belong into the second group.

My point here is that the vast majority of money spent in the OSM community (both by private businesses and non-profit organizations, both directly in the form of paid time and indirectly in the form of subsidies and investments) is targeted at the third group. But what people often don’t realize is that this focus on the third group is

• interpreted by members of the first and second group as a sign of lack of appreciation for their intrinsically motivated volunteer work.
• leading members of the second group to orient themselves professionally away from OpenStreetMap, because of the lack of demand for their work compared to that of the third group.
• incentivizing members of the second group to change towards working more like the third group and let their free time involvement be more influenced by economic forces.
• putting pressure on members of the fourth group to get involved in OpenStreetMap beyond their paid time.

Some of these effects might be considered desirable. But keep in mind what i said earlier: That a huge part of the human potential in the OSM community exists in the form of the first two groups. It is also my observation that the third group tends to be the most homogeneous of all four groups – most male dominated, least culturally diverse.

At the Karlsruhe Hack Weekend there are people from the first three groups present (the fourth rarely so – due to the nature of the event) and because of that it can function as a forum of exchange between these groups. But you can also feel the economic and social dominance of the third group, which sometimes makes a balanced exchange of views and ideas difficult.