Check out this thing which I just got working again:
>>> Long Names of OpenStreetMap <<<
So that’s elements with a name tag, where the name seems to be very long. It’s a full list of the longest names in the planet (>150 chars)
I just found this old code and dusted it off. I made it originally back in May 2010 when Richard Weait ran a “Project of the week” looking at long names. I called it “namecheck” at the time, but I’m renaming it “longnames”.
A crazy long name is probably a data bug to be fixed. It certainly looks like a bug when we try to render it.
But… well taking that example. Here we have a few university buildings, each with a name, which list several departments. Can we class this as wrong? I think so, yes, but I’ll have to make a hand-wavy non-scientific judgement: I’d say it ceases to be a name of the building and becomes more of a “description” when it gets that long…. in this case. And maybe that’s a problem for a lot of them.
But in general this is not going to be a clear case of “these are all bugs. Here’s how to fix them”. Because of that I think I would stop short of saying this could be a MapRoulette challenge, or a new class of bugs for KeepRight or Osmose. For the moment I’m just going to put this list out there, and suggest some careful interpretation (but feel free to suggest other actions / tagging policies in the comments)
On the other hand some of these definitely are bugs with clear fixes. I saw somebody adding a load of space characters and then repeating the name again within the name tag, presumably in attempt to control label placement on a road (Yeah no we don’t do that)
The other caveat, in case it wasn’t clear, the username listed there is ‘last editing user’. It’s not necessarily the person who added the very long name, but it is a person who didn’t fix it :-) That and the last edit timestamp are really just to see if there’s any interesting patterns in the list.
My planet crunching script to find long names is here along with SQLite, CSV, and mangled OSM XML output files. I will re-run it again in a week, but the display also has a (slightly odd) feature to let you “rescan” an element after you’ve made a fix.
[Update: I mentioned “feel free to suggest other actions / tagging policies in the comments”. But also this discussion on the wiki might be a good place to discuss debate what the policies should be]