OpenStreetMap

Being a newbie

Posted by escada on 21 November 2014 in English (English)

This is story is based on real stories. It is not my story as a newbie, but I decided to write in the first person to avoid she/he discussions. Also, since English is not my native language, so I apologise upfront for mistakes.

I love to ride by bicycle and for plannng my trips I found those great free maps offered by OpenFietsMap. I used them during my vacation in The Netherlands and now I want to improve the map for cyclists in my hometown in Belgium.

I created an account on OpenStreetMap and quickly found out how I could launch the iD-editor. It seems pretty simple to add a separate cycleway, just as I saw on the map in The Netherlands. I think it is important to see the difference between street with and without those separate cycleways. So let's try to add them.

O great, there are arial images that I can use, so I do not have to upload tracks that I recorded with my GPS. OK, let's see, the cycleway starts here, in front of those houses. So I start drawing the line there and continue here, cross the street and it ends here in front of this parking lot. Now add some tags to it...mmm .. a name... mmm maybe "fietspad" (Cycleway in Dutch).

Ok, now the other side. Mmm, the houses that the previous mapper placed are on top of the cycleway. I'll move them so I can draw the cycleway in the correct place.

Hey, that was easy, let's save this so the others can enjoy my work. O, do I need to add a comment... mmm ... "Fietspad" will be ok I hope.

So far the first editing session from an newbie user as I see it. The user honestly tried to improve the map. But could you spot some mistakes ? Here are some

  • the cycleway is not connected at start or end
  • The cycleway has no intersection with the street that it crosses
  • It's tagged with a name that indicates its function
  • the user did not put bicycle=use_sidepath on the main street
  • the user did not remove any cycleway= from the main street
  • the user is unaware of relations for cycle routes on the main street that have to be placed (and splitted for the different directions) on the cycleways
  • the user did not add oneway=yes on the cycleway
  • using Bing images which have an offset, in Flandres we can use AGIV, much better

Not all of those mistakes are made by all newbies and maybe they make some I forgot to mention here. But that is not important for the message I want to bring. One can make many mistakes and none of the editors protect you from making no errors. Some editors protect you from some of the above errors, but many mistakes pass unnoticed.

But now dear experienced mapper,

How do you react when this happens in your neighborhood ?

Do you

  • yell "vandalism" ?
  • contact the DWG ?
  • start complaining on a local maling list about this user that destroys all this hard work ?
  • send a angry private message or changeset comment ?
  • do you ally with your friends to send multiple scaring changeset comments ?

or do you take a deep breath, relax and try to write a friendly, polite message to help this newbie navigate through all the pitfalls and unwritten rules from which the editors do not protect you ? Even if you have to do this for the tenth time ?

So think for a moment how it feels to be a newbie and receive a message from some stranger about something you honestly thought was a good addition to OpenStreetMap, next time you write a comment about someone else work. Heck, even when that person is a more advanced mapper.

Happy mapping & communicating

p.s. I fear that the real story that was the basis for this one does not have an happy ending

Comment from Richard on 21 November 2014 at 13:16

Brilliant post. Spot on.

We probably need to make (future) private messages into public messages to prevent this form of bullying, and provide a feed of all changeset comments so that bullies can easily be spotted there.

Comment from escada on 21 November 2014 at 14:17

Thanks Richard.

Seems like a good idea, but private messages between "friends" should remain possible.

Comment from JBacc1 on 21 November 2014 at 18:20

And still the same question… Is it better that the newbie does not know he breaks things when saving (and is not afraid of going on), or that he gets some warning message saying, hey, check this, it looks strange. My opinion is clear-cut, but it is thought that iD « must not scare newcomers ».

That's why I love JOSM, even for the first time they edit! And that I use it in carto-parties.

Comment from escada on 21 November 2014 at 18:39

Hallo JBacc1,

What I try to communicate is not whether the editor should warn about everything, but that the community should try to act in a friendly way towards newcomers every time again, just because the editors do no prevent one from making mistakes.

Note that JOSM will not warn you that you did not place a bicycle=use_sidepath on the main road, or redirected the relations over the cyclepath. It might catch more stuff than iD. I don't know, I hardly ever used iD. I started with JOSM from day one.

Comment from Zverik on 24 November 2014 at 21:55

Thanks for this story, I've translated it into Russian. We have a bit of a problem with angry reactions.

Comment from joost schouppe on 27 November 2014 at 00:41

The solution of course can't be all technical, but I do believe editors should warn you when they think you're doing something strange. But how about giving a contributor the option to "send changeset for review" when this happens? An experienced user can probably approve, comment or disapprove ten changesets in the time of one revert.

This is a recurring issue on the diary and on the forum. I think we should try to join forces and write up some proposals. I'm sure that we can do better than we are doing now, both protecting data integrity and contributors making mistakes... Then, it will be easier to stay patient.

Comment from joost schouppe on 27 November 2014 at 00:56

Examples: The sad story of a new contributor messing up http://forum.openstreetmap.org/viewtopic.php?id=27704 , and someone who's annoyed http://www.openstreetmap.org/user/Gerald%20Weber/diary/23680

Comment from escada on 27 November 2014 at 07:52

@Zverik: I hope it has some effect for your community. @Joost: The idea of a review is certainly worthwhile to investigate. For me such a review should be done by someone that speaks the same language as the submitter, moreover, it should be someone that is as local as possible (same town, if not, same province, if not same country). This might pose some problems for small communities. Where do you find the volunteers that want to review changesets in detail ?

How do you become part of those "experts" ? I've got some advice in the past from self-proclaimed experts that turned out to be not so good. Not everybody with hundreds of changesets is an expert, or an expert in all fields.

You also want to avoid that a bully reviews those changesets.

Looking forward to see a first proposal from you. I rather spend more time on mapping than writing guidelines, proposals, etc. :-)

Comment from joost schouppe on 28 November 2014 at 21:09

Just passing me the ball now, aren't you :)

Well, I do like to think about things like this, and I don't have enough internet at the moment do do a lot of mapping (apart from making notes and adding POIs). So maybe, maybe.

Comment from tmcw on 2 December 2014 at 05:22

Great post, definitely highlights the core problem here: beyond rules and validation and so on, there's a social question and a question of onboarding.

I'm optimistic about comments on changesets. If they're friendly, then experienced users could, for instance, review some percentage of new edits and create & explain their corrections in the same sense as the editor of a book or how code reviews work in programming.

If some voluntary review process, like a checkbox to say "I'm not sure about this, can someone check my work", were technically feasible, it could be a good thing. I'm not sure that is technically feasible at the moment, because of the infrastructure required and the increased chance of conflicts. But, it would be a fun thing for someone with time and energy to experiment with.

Good real-world systems are error-correcting and human.

Comment from escada on 2 December 2014 at 06:15

Thanks for your comment tmcw.

Changeset comments can be great, but I recently witnessed a case where several people started commenting on the same change. And while each individual comment might be friendly, just the amount of comments might be intimidating as well. (the ally with your friends case)

From time to time you see this request for comment made by people, either on a mailing list, the help-website or a forum. Of course the change is already made in such a case.

It's my impression that the experienced mappers are in general friendlier towards people that say "hey I made this change, can you check it?". So offering this functionality from within the editor would be really nice. Still the question remains, who are the "experts" ? What in case of different opinions of tagging something ?

Comment from joost schouppe on 3 December 2014 at 17:25

Anyone who defines themselves as an expert on OSM is an expert on OSM. Do we need to formalize that? If we do, we might restructure our communications system Reddit style. Edits and comments can get you positive and negative "karma", which is the opinion of the rest of the community. Bullies would easily get downvotes, making their behavior less rewarding.

It's the nature of OSM that opinions will differ. New contributors who cannot deal with that, might not be the most suitable contributors anyway. But the more discussion about "the right way to map", the more probable that some formalization will take place, right? If you have a real voluntary "human validation process", you do indeed need some guidelines. As @tmcw said, database conflicts might be a problem. So approval has to be as quick as possible. Changeset approval and discussion should thus be separated. If things are "about right", approve the changeset and open the discussion afterwards on how things can be done even better. Only if things are really bad (breaking things, doing things in completely different ways from standard practice), disapprove. For example, breaking a relationship would result in disapproval, mapping landuse including ways would be approved but commented upon.

Again, anyone could sign up to approve or disapprove changesets. But anyone can "upvote" or "downvote" their decisions. People with too many downvotes could be excluded from the process if the need arises.

When it comes to time and energy, I believe the first priority should be on making some automatic error detection. Because even experienced users make mistakes, automatic error detection can help everyone.

Comment from KrystleChung on 25 December 2014 at 02:37

This was interesting to read. I'm in the community support role at another crowdsourcing project (wikiHow.com) and this is one of the main things I help to mediate. Newcomers make the same mistakes and ask the same questions again and again; it can be hard for long-term contributors to not get burnt out by this because it feels repetitive, even though it's a different person every time, and each person is often innocently asking a reasonable question, or making a common mistake. I'm curious to learn more about how OSM deals with this.

Login to leave a comment