Recent diary entries
It is the youngest capital in Europe, a city where years of peace have brought constant street name changes and endless new streets being built. Locals often use landmarks – sometimes long gone – or old or colloquial names when giving directions, which I had to ask for repeatedly on my first trip to Kosovo in May to visit Anne. It's a cliché to talk of Kosovo’s potential, but it felt especially true when I looked at the map of Pristina, trying to find my way from Bill Klinton Boulevard, which has a huge statue of him, to Rruga B, which has changed to a name no one remembers or uses.
Big proprietary map makers can’t and don’t care to keep up. Even more than elsewhere, the best map of Kosovo can only come from local mappers.
On my next visit, together with the wonderful OpenStreetMap community, I organised a meetup. This led to very productive meetings with Kosovo’s national cadastre agency. I’ll hopefully have more to write about open geographical data in Kosovo soon.
Other meetings happened with the city of Pristina, and Yll Rugova, who is among others the director for tourism for Pristina, showed up at our second meetup a week ago. We talked about the challenges that come with putting Kosovo on the map, and his ongoing project to create a tourist map of the city. He looked delighted when I showed him Maposmatic, and how easy it was to generate what he needed in different sizes and styles and modify the PDF output.
Only a few days later, this is the result (or in Albanian) – the official tourist map of Pristina is based on OpenStreetMap. The translations and printed versions are coming soon, and I'm sure other, even better maps will follow. Congratulations to Yll, to the Kosovo mapping community and to the city of Pristina.
(Edited to add link to English map)
I will attend it to present our guidelines and answer questions. In the meanwhile, here is my report. I'm also happy to answer any comments below.
We at the DWG are, first of all, thankful for all the constructive input we have received, from the advisory board, the humanitarian mapping initiatives and the mapping community.
This version of the organised editing guidelines took a lot of work to prepare. We have received and integrated a lot of feedback to reflect consensus and existing good practice.
We have looked at what similar policies would exist, on OSM or in other organisations. I believe that no other project, open or proprietary, has faced this exact issue before. On OSM, contributors generally understand the current policies on automated edits and imports. We have written the organised editing guidelines in a similar way, while adopting a slightly softer approach – not following the organised editing guidelines isn’t an offence per se. Elsewhere, Wikipedia has numerous policies some vaguely similar, but the problems they face are quite different, and their policies tend to be a lot more complex.
Internally, we have looked back at past problematic edits. We have carefully written the guidelines and defined the scope to prevent those problems without creating loopholes or negative incentives like encouraging salami tactics. They are not meant to apply to community activities like mapping parties between friends or making a presentation on OSM at a local club, but only to ‘sizeable, substantial’ activities. We want something that doesn’t scare casual events off while letting us regulate a geography class gone berserk or a misguided volunteer mapathon.
We also don’t want to set hard limits in stone since they would have to go back to the Board constantly if we need to refine exactly what falls under the guidelines.
Humanitarian activities deserve our fullest support. We have therefore adapted the guidelines for them, both implicitly, by requiring only a best-effort approach, and explicitly, by exempting emergencies from the two-week discussion period. Some humanitarian edits have been problematic before, and the guidelines are easy to follow; a blanket exemption would send the wrong signal.
We see the amount of corporate good will at SotM, the tensions in the community, and the (dis)organised edits that mappers refer to us. Publishing the guidelines will be good for everyone. Good actors, existing and new, will be able to trust clear expectations. The community will be confident that this is the consensus that will be respected. Confused newcomers will get a blueprint for a successful organised edit.
We have written guidelines that will be easy to read and follow and provide clarity on how good organised edits should run without having a chilling effect on them. DWG proposes that those guidelines should now be officially published. OpenStreetMap needs them now more than ever.
Inspired by Jochen Topf's old style multipolygon writing, I have just finished manually cleaning up (most? all of?) the remaining old style MPs in Luxembourg.
The process is currently quite tedious. If we want to put old-style multipolygons behind us, we need to make it easier to find them, convert them, maybe even get editors to automatically convert them on the fly when they're touched. QA tools like Osmose could also possibly make it a one-click JOSM remote fix.
The method I used was downloading a Luxembourg extract, then searching for
type:relation type=multipolygon tags:1 in JOSM. I also had
-((type:relation tags:1) | (child (type:relation tags:1))) set up as a filter, but didn't end up using it, because the MPs would disappear as soon as I'd tag them.
For each MP I found, I would zoom to it (
3), click on the outer way, alt-click on that way to select the relation, copy tags from previous selection (utilsplugin2,
shift-R) from the way to the polygon, click the way again, and delete the redundant tags on the way.
Sometimes, for joined geometries, I had to select the relation by right-clicking on the parent relation in the tags list.
Since most of these old-style relations haven't been touched for a long time, I also did some drive-by cleanups. In Luxembourg, the old-style MPs were mostly forests.
Did I get all old-style MPs, or only most of them? If there are multipolygons out there on which both the outer way and the relation are tagged, my find pattern didn't find them. These are possibly the most ambiguous ones, e.g. conflicting tags on outer way and relation, and should be fixed. Has anyone got a good easy way to find them? I tried
hasRole:outer -untagged but that found a lot of false positives.