Amᵃᵖanda | map data witchThe #OpenStreetMap equivalent is Craft Mapping by Craft Mappers FYI.
Taggart :donor: (@[email protected])
Hello! Are you interested in artisanal, handcrafted software lovingly assembled by master programmers? You will be. *You will be.*
Infosec Exchange
“Craft Mapper” originated ~10 years ago. One of the American OSM Facebook dudes wrote that #OSM was doomed unless unless big tech used AI to save it from the craft mappers, who would drive the project to oblivion in 5-10 years.
So people (incl. me) gleefully adopted Craft Mapper as something to be proud of.
The #OpenStreetMap project has been dealing with #AI #slop for almost a decade. It's sooooo bad. It's so rubbish. The quality is bad. It never meets what the salesmen promise. It takes much more time to fix it up than it would have taken to do it properly.
Mat
Manuel 🦡🦡 ⚾@okwithmydecay @amapanda @rapid true, but they are traced from Bing Aerial Imagery which can lead to weird or bad suggestions
@okwithmydecay Facebook has made the whole Rapid development team redundant. It works now, but eventually will get worse. It's not being maintained or developed anymore.
Iván Sánchez Ortega@amapanda Surely if big tech would have foreseen this they'd form a OSMF-like consortium to support a AI-derived map data project that would *totally* have displaced OSM by now, right?
buwel
Marcos Dione@amapanda anyone selling stickers?
@mdione there are some craftmapper stickers. I think we have some in Geofabrik. Maybe some might be at SotMEU.
There are craftmapper t-shirts at FOSSGIS events
Bastian Greshake Tzovaras
Dorothea@mdione not what you asked, but you can also print those stickers yourself or create a t-shirt, if you wish. @amapanda had added the design to the OSM wiki: https://wiki.openstreetmap.org/wiki/File:Craft_Mapper.png
Lunaphied@amapanda wait what was the supposed threat from craft mappers?? just that they wouldn't be AI? also lmao you can't use AI to collect map data that's literally not possible
@Lunaphied you could in theory use AI to trace buildings and roads form aerial imagery.
In practice, it doesn't work anywhere near as well as you need for a map. Which is why we don't use it.
We first found out about Facebook's stealth import of AI derived roads because it broke the routing graph.
@Lunaphied the Facebook dude said that if those old cranky craft mappers didn't get out of the way and let FB's AI slop in, the OSM project wouldn't be able to keep up.
You can read the post here http://mike.teczno.com/notes/openstreetmap-at-a-crossroads.html
Eugene Alvin Villar 🇵🇭Tagging the author just so he has the opportunity to respond: @migurski
Michal Migurski 🍉@seav @amapanda @Lunaphied I’m not going to super get into it because Amanda links my original post above, but my observation at that time was that the northern European communities were fixated on craft methods that could not be reapplied to where OSM needed to grow, and hostile to methods that could be including HOT’s work. I wasn’t an “FB dude” when I wrote it, that job came later 2018-2022 and I followed up with a lot more written arguments in a handful of board campaigns.
@seav @amapanda @Lunaphied …what we did at FB was to set up the Overture Maps Foundation so if you’re interested in what specifically the alternative vision was, OMF is the right place to look. In 2016 I was merely frustrated, by ~2023 we had a coherent response organized inside the Linux Foundation.
@seav @amapanda @Lunaphied Sadly (for me, maybe not for you) my roles since leaving FB/Meta haven’t really given me much of an opportunity to focus on OSM politics so I’ve pulled way back to just OSMUS events and trying to keep my OSMF membership current
@migurski I always find it sad when less people are involved in OSM politics/activity/community stuff. I want more people active.
@seav @Lunaphied
@seav @Lunaphied
@migurski “my observation at that time was that the northern European communities were fixated on craft methods that could not be reapplied to where OSM needed to grow, and hostile to methods that could be including HOT’s work”
Here we are ~8 years later. What's your opinion of how that paned out? Where are we now?
@amapanda @seav @Lunaphied My opinion is a couple years out of date so take with a grain of salt but I see two really big changes in the past eight years. A positive one is that OSMF has paid technical staff vs. the old pure-volunteerism approach, so there’s more sense that problems can be solved through normal work. A mixed one is that the OSMF seems to have kept focus entirely on methods and no longer on outcomes; it’s Overture & HOT who care about map growth and utility now.
@amapanda @seav @Lunaphied …which was the source of my frustration in 2016: the original team started OSM with the idea that an outcome (global map) could be achieved through a method (individuals mapping locally). The vocal community later (ca. 2009) chose to privilege the method over the outcome even as it started looking like the method couldn’t simply transfer outside of Europe without adaptation. There’s still a general vibe of “you’re doing it wrong” directed at Overture, HOT, etc.
@amapanda @seav @Lunaphied …but the political situation is fairly stable now. If you care about the method you can focus your attention on OSMF for get-togethers, opinions on feature tagging, mapping parties, etc. If you care about the outcome you’ve got Overture or HOT to talk to for expanded geographic coverage, certainty about data consistency, and usefulness of map data for a purpose.
I’ll close by saying again that my opinion is no longer informed by an international OSM perspective.
Rihards Olups@migurski @amapanda @seav @Lunaphied I'm still getting an impression that Overture is more about wrestling control out of the hands of everybody else in favour of big corporations that can the control all the critical paths.
All under very practical and not even outright wrong reasons - but the outcome is something that is much more prone to enshittification by suffocating OSM.
We've seen that end up bad for everybody but a very, very small number of people - and we still keep on falling for it.
All under very practical and not even outright wrong reasons - but the outcome is something that is much more prone to enshittification by suffocating OSM.
We've seen that end up bad for everybody but a very, very small number of people - and we still keep on falling for it.
@richlv I see what you’re saying but I have a different interpretation. Commercial founders of Overture tried working on the global map outcome with OSMF directly and openly but got rejected. They still needed the outcome so they then decided to work separately via a new organization. Nobody at any point wanted to suffocate OSM. The individual mapping method is simply uninteresting and irrelevant to the outcome, not worth controlling, enshittifying, or thinking about.
@amapanda @seav @Lunaphied
@amapanda @seav @Lunaphied
@migurski I disagree that there was an emphasis about the method. Me importing gov't data back in 2006 was more than welcomed.
@IvanSanchez @amapanda @seav @Lunaphied I think the method focus came later: when I got involved in 2006 there seemed to be a lot of flexibility toward discovering other methods and welcome space for activities like the TIGER import. By 2-3 years later that no longer existed. Somewhere in there was an acrimonious license fight as well, but in retrospect I don’t actually think the CC-to-ODBL change was the meaningful one.