emacsen's diary

Recent diary entries

Why SOTM EU was awesome

Posted by emacsen on 22 June 2014 in English (English)

Days like today...

Posted by emacsen on 28 April 2014 in English (English)

Today has been an incredibly frustrating day in OSM for me.

I started off the morning dealing with DWG stuff. There's so much negativity in some people, but when communicating on behalf of the DWG, one must try to rise above the petty insults and most past it. Then I had to deal with some domestic mapping issues before they rose to the level of DWG intervention.

Then I ran a mapping party, where for the second time in a row, only one other OSM NYC member showed up, despite us getting 6-7 RSVPs. I spent a lot of time prepping for the meeting, finding the right place to have it, finding the right place to meet, etc., and when people don't show up, it's very disheartening.

Then I find that NYC has had dozens, maybe hundreds of POIs messed up by an automated edit, by a user who was clearly told to stop his automated edits.

And then I get an email complaining about bot-mode's TIGER expansion project, a project which I spent months working on.

Then, finally, MapRoulette had a new contributor submit a patch, a very long patch, which clearly took them a long time to write, but I had to reject it because it didn't fit in with the bug that it was supposed to solve, nor did it fit generally in with our programming style/the way we organize the code in the project. It just would not have fit. I tried my best to give honest, supportive feedback, but I know that rejection always stings, no matter how it's presented.

Days like today put me in a bit of an OSM funk

What kind of user/challenge stats do MapRoulette users want?

Posted by emacsen on 18 April 2014 in English (English)

Last night I was working on the statistics display system for MapRoulette and realized that I don't know what to display.

In the rewrite of MapRoulette, we knew that me trics would play a large role in the project, so we added in lots of statistic capturing capability, as well as piwik support for even more metrics.

We could track what challenge a user has worked on, where those challenge are, where specific tasks are, how difficult the challenges are, how a user classifies tasks, how long they spend on maproulette, even how long they spend (generally) on each task.

Similarly, we could collect a ton of metrics for challenges. How many users work on it, are there a few "super-users" or is it widely distributed? Are people saying the tasks are fixed, skipped or false positive, etc. We can even find out how long users spend in aggregate on the challenge, or even drill down to the tasks and find out when people walk away from them.

But amongst all the possibilities- what do MapRoulette users really want to see? What intouerests them? Stats about themselves? Stats about other users? Stats about challenges?

Tell me what kind of stats (if any) you'd like to see in MapRoulette?

The Maintenence of Imported Data in OpenStreetMap

Posted by emacsen on 13 March 2014 in English (English)

I've written a new blog post, The Maintenence of Imported Data in OpenStreetMap:

Video interview with me about OSM

Posted by emacsen on 30 January 2014 in English (English)

Why Imports in OpenStreetMap Are Controversial

Posted by emacsen on 26 January 2014 in English (English)

Edit Wars in OpenStreetMap

Posted by emacsen on 18 January 2014 in English (English)

Why the world needs OpenStreetMap

Posted by emacsen on 5 January 2014 in English (English)

I wrote a post on my blog entitled Why the world needs OpenStreetMap

Motivation for Contributing to OSM

Posted by emacsen on 6 December 2013 in English (English)

The issue of motivation of mappers has come up a lot lately in different contexts, be it commercial interests and imports, adding specific POIs because they share a common payment form, adding notes that advertise a business, etc.

Ideally, we wouldn't care about why someone is motivated to contribute to OSM- if the data is good, we accept it, and if the data is not good, we don't. Unfortunately the reality is more nuanced than this. The reason that someone wants to add data to OSM can significantly influence the way that they enter the data.

Let's take the extreme example first about notes that are meant to advertise a product, this note here:

That note simply says:

We are proud to present our new Search engine

Now, I don't think the motivation of this person is very good, and in the meantime the account has been deleted for spam.

But we have more subtle examples of adding businesses to OSM, as shown by our discussion about addresses of offices vs mailboxes.

In regards to imports, there have been concerns that importers may be responding to commercial pressure to get a certain dataset into OSM. So far those concerns have not been validated by facts, but the concerns are real.

I think this question of motivation is at the heart of many of the recent disagreements. Two mappers may have a disupute and that's acceptable. It's also acceptable for a new mapper to come in and make mistakes (I certainly made plenty in my first few months in OSM). But where our community is less tolerant of error is where they feel the motivation of the mapper is poor.

The other reason this is important is that poorly motivated efforts shift the work back to the community. The community is build on improving others work, but when the work is done without caring about the end result, the burden on the OSM community is shifted and the community has to "take up the slack", which moves us from being contributors to cleaning up others' messes.

What can be done about bad user motivation? Or is there anything that can be done? Do we need to raise the bar on new user contributions, or is this just a cost of being a successful project?

Older Entries | Newer Entries