woodpeck has commented on the following diary entries

Post When Comment
Are maproulette challenges undiscussed mechanical edits? about 2 months ago

I agree that we do have to think about how to deal with large-scale directed mapping activity, and what requirements we have there. Like with an import or a mechanical edit, a mistake that someone makes in designing a task for 100 employees or 100 students or 100 contractors to follow, can have dire consequences and they can unfold quickly. I think that Martijn's plan to have volunteer challenge administrators is a good first step provided we can find the volunteers ;) community guidelines about directed mapping can follow later.

Are maproulette challenges undiscussed mechanical edits? about 2 months ago

Further accounts with similar patterns: * * * * * * * * * *

New data sources available for Western Australian roads 5 months ago

@Sam Wilson, +1 to your considerate use of that external data source to improve OSM. It is a pleasure to see someone choosing the manual, careful path rather than just throwing together a half-baked script that tries to automatically import the lot and breaks half of the data while doing so.

Share with Twitter / Facebook 5 months ago

Share with Flickr, Telegram, Snapchat, Instagram, Twitter, Slack, WeChat, Vkontakte, Tout, Google Plus, WhatsApp, share on LinkedIn or Pinterest, or any other of the 25 social startups we're seeing every month, after all we don't want to apply any bias!

Umfrageplattform offline 7 months ago

Was ich über Schiedsrichter gesagt habe, gilt auch dann, wenn die Schiedsrichter mal mit verschiedenerlei Maß messen. -- In der DWG haben wir das recht häufig, dass wir jemanden wegen etwas ermahnen, und die beleidigte Reaktion darauf ist, dass diese Person uns künftig jeden anderen meldet, der (ihrer Ansicht nach) das gleiche tut. Das ist aber kindisch, denn bei einer Ermahnung kommt es auf eine Vielzahl von Faktoren an. Es gibt nicht strikt eine Regel und jeder, der dagegen verstößt, wird ermahnt.

Beispielsweise erwarten wir oft von einem, der länger dabei ist und ein erfahrenes/angesehenes Mitglied der Community, dass er sich besser verhält als ein Newbie, der es vielleicht nicht besser weiss. Ich finde es völlig normal, von einem alten Hasen wie Dir - der eben durch Erfahrung und Alter auch ein Ansehen und eine Vorbildfunktion in der Community hat - eben auch ein vorbildlicheres Verhalten zu erwarten als von einem Neuling, der verzweifelt "finde meinen post nicht" schreibt.

Dein im letzten Absatz geäußerter Plan, das Forum mit einem gespielt einfach gestrickten Gerechtigkeitsempfinden absichtlich zu stören, wird sicherlich das erhoffte Resultat bringen. Damit hast Du dann allerdings niemandem etwas bewiesen außer Dir selbst den eigenen Dick- bzw. Hitzkopf. Hast Du das nötig?

Umfrageplattform offline 7 months ago

Bist Du tatsächlich im Forum gesperrt worden, oder ist Dir nur eine Sperre angedroht worden ( Falls Du gesperrt worden bist - ist der Grund dafür, dass Du nach der Androhung noch weitere nach Moderatorensicht provzierende Postings geschrieben hast? Falls Du das getan hast, warum hast Du das getan? Sind Moderatoren Deiner Ansicht nach unnötig, oder sind bestimmte Benutzer irgendwie zu wichtig, als dass sie auf Moderatoren hören müssen?

Mit Moderatoren ist das wie mit Schiedsrichtern. Wenn die was entscheiden, muss man sich dran halten, auch wenn man es unfair findet, sonst eskaliert die Sache. Man kann abwarten, bis das Spiel vorüber ist, und dann in Ruhe diskutieren, ob der Moderator vielleicht ungeeignet ist oder die Regeln, nach denen er arbeitet, geändert werden müssen. Aber in der Hitze des Gefechts ist eine solche Diskussion nicht möglich.

On Being Trumped 7 months ago

Over here in Old Europe where people tend to have a residual respect for everything that is "official", it often helps to wear a high-viz vest emblazoned with "SURVEY" or so. Certainly more conspicuous, but even less suspicious than a suit-and-tie ;)

New script to convert SHP into OSM relation boundaries ( 8 months ago

It appears the embedded images don't load. There's a 6-year old script that seems to do similar things without involving QGIS here: - I'm not sure what the differences are in detail. I would also encourage you to be a bit more verbose when you say: "Of course, make sure to clean up whatever boundaries were in the area you want to upload" - it is certainly not acceptable for someone to simply delete the already-existing boundaries to replace them with an upload, so the correct workflow would have to include downloading the existing data and comparing it to the data from the SHP files that you want to work with. You do point to the import guidelines but there's a danger of someone stumbling over you script on GitHub and thinking they can just run it without further ado; perhaps you could add a word of caution there in a README or so.

Super-broad "self-explanatory" tags 8 months ago

But is the "problem" you point out not a general one with OSM? For example, in Iceland you can expect practically any petrol station to have soup and bread available in case you're hungry. I'm pretty sure this is not the case for petrol stations world wide. But will the Icelandic traveller be so clueless that we actually have to use special tags for this (amenity=fuel, soup=yes in Iceland, and amenity=fuel, soup=no in Argentina)? A shop where you buy alcoholic beverages is totally unregulated in some countries, but will be state-owned and tightly regulated in Sweden and Norway. A bar in Italy will sell you a croissant and a cafe for breakfast, wheres a bar elsewhere will not open before nightfall.

OSM is a global database, but do we really have to equalize everything - do our tags really have to describe the reality to someone who has zero familiarity with the area in question?

Translating places with CityNamer + Bots 8 months ago

Transliterated or transcribed names should not be added to OSM unless there is some human expertise required to select the correct transliteration or transcription. If something can be done automatically then there's litte use in adding it - it would be like adding a "length" tag to a street, adding no information and risking bitrot when the street is modified.

Names should never be translated as they are likely to lose their property of being a name (e.g. the Pont Neuf in Paris translates as "New Bridge" but that is not its name in English).

Many things have genuine names in different languages (e.g. München, Munich, Monaco di Baviera, ...). These are currently accepted in OSM but my personal hope is that we can shift this job to Wikidata because I certainly do not enjoy the prospect of having a tag inflation like for every named thing in OSM.

Introducing OpenStreetView 9 months ago

I'd say an imagery platform like this is really open if it lets me do things like: * write a bot that analyses all imagery in a country and tries to auto-detect which roofing material is prevalent in which area * test a new street sign recognition algorithm against all images that are taken along streets * find images of this area that have snow on them * fork the whole thing and run it myself because I don't like the guy who runs it

I'm not asking for server capacity that implements these things for me, and I understand that making terabytes of imagery available for anyone who asks is a challenge, but I need the data to be free and open and accessible enough to allow these use cases. If the provider of the service reserves these use cases to themselves and installs themselves as a gatekeeper between me and the data then the platform might still be a great tool (like e.g. the Bing aerial imagery is for OSM) but certainly not open.

What, if anything, should I do about this? 9 months ago

Even though it is difficult, I'd recommend simply leaving things. The person not wanting something in OSM is going to lose in the long run because every mapper who walks down that track will add the building again. It's no use getting all worked up about it - in only reduces the happiness of at least two people.

Adding columns to a non-updatable import. 10 months ago

Another approach would have been to filter out all objects that have a castle_type column using osmosis, and then feeding the resulting file to the database via osm2pgsql -a (provided you have done your original import with --slim).

Note that your approach, as documented, does not work for relations as those will appear with negative IDs in planet_osm_polygon; you need to duplicate the "way" method and do a "... where osm_id = -%s" for this to work.

Mapping Wikipedia 10 months ago

Depends on where you inspected the additions. The OSM web map is quick to update but doesn't have clickable links to any URLs for fear of being abused for link spam. Other maps or apps might enable such linking but may only update from snapshots.

OSM - beschissene Datenqualität 11 months ago

Es gibt viele, die dem "OSM-Datenchaos" begegnen, ohne dabei anzuecken. Wenn Du das nur mit Problemen kannst, dann ist es vielleicht am besten, wenn Du es bleiben lässt. Danke für die Einsicht.

Group of buildings imported in Boscoreale, Italy 11 months ago

Imports have to follow the import guidelines, most notably, they have to be discussed before they are executed. The usual remedy for a non-discussed import is not to open a Mapbox ticket to have it cleaned up, but to revert it altogether and then give the community a chance to discuss whether they even want this data, whether license is ok, the data quality is good, and so on.

The importance of sidewalks about 1 year ago

I'd like to point out that better pedestrian routing is possible without drawing footways individually; see and the (sadly, German) Bachelor thesis linked from there. This is about a piece of software that generates a pedestrian routing graph from OSM data, by synthesizing pavements/sidewalks as individual edges. Pedestrians will therefore not be routed along the road centreline, but along the left or right side.

Why local assumptions are wrong for an international project about 1 year ago

This "problem" has existed in OSM for as long as it has been used internationally. For example, both Germany and France have a distinction between a bakery where you mainly get bread, and a bakery where you mainly get sweet pastries and cake (Bäckerei/boulangerie and Konditorei/patisserie). Both are shop=bakery in OSM. In Italy, people visit a bar for their morning croissant and coffee, while in Germany a bar won't usually open before 20h. I'm not sure but I think both are amenity=bar too. In Iceland, practically any petrol station has a small eatery attached where you can at least get a soup and bread - this doesn't have to be mapped explicitly. In other countries this might be a rare exception.

You can't hammer your computer science thinking into everyone. You can't get Italians to stop calling their bars bars just because they don't fall into your bland international definition of "bar". You can't tell Icelanders to add "food=yes" to all their petrol stations just because it wouldn't occur to someone unfamiliar with the country. I think that a little local flavour isn't a problem for us, and we shouldn't try to level all these nice national intricacies.

I think that in the long run, we'll have to have national rendering rules, so that the communities in various countries can decide what is usually shown on their maps and when, and what kind of icon is used. And yes, perhaps in the long run, when every building is tagged with how many seats it has and whether food is sold and what kind and when (in addition to the roof shape and the heating technology of course), then you might even be able to let your computer find out, culture-ignorant, where you can go to buy a slice of cake. But until then it probably depends on the country you're in whether the shop=bakery around the corner is likely to sell any.

Welcome to the new Missing Maps about 1 year ago

If I strip away the "this helps the vulnerable people" bit and just look at this from the OSM side, what remains is: "An external organisation creates, without discussing this with the OSM community on any meaningful level, a system of organising and rewarding volunteers to contribute data of that organisation's choosing to OSM, in a way and level of detail controlled exclusively by that organisation. The external organisation massively outspends OSM in publicity, has software developed professionally, and thinks of volunteers primarily as 'their volunteers', not OSM mappers."

I think that this sets a dangerous precedent and OSM should not blindly allow such organised influence on what goes into OSM and how. There needs to be a consultation process because otherwise we'll have lots of other interest groups queuing up to send their volunteers (or rewarded volunteers, or employees) our way to map whatever that interest group wants to have mapped.

Also, rewarded mapping activity bears the danger of encouraging over-eager contributions that are precisely geared to maximum reward and not maximum usefulness for OSM. Problems like that have surfaced with nearly every gamified/rewarded mapping in the past, and Missing Maps is not an exception.

We must be very careful not to allow Missing Maps, just because "it is for a good cause", to blaze a trail that others will want to follow.

PS: Is Missing Maps open to the degree that we get to see how much money they spend on what, and who gets to make the decisions? As a member of the OSM Foundation, if some group prides itself in that level of influence on OSM, I'd actually like to know who exactly I am dealing with.

I get why so many people just don't bother to even try to vote for the OSMF when things like this happen over 1 year ago

I was contacted by a web IRC user during the meeting when the servers were down, saying that they could not login. I explained to them that at that point, voting had already closed for ~ 30 minutes (voting was closed early in the meeting). The voting was closed when the servers were still fully working. I'm sorry if it was not made clear enough that voting would only be possible until the meeting, not during all of December 5.