[Talk-us] Community Involvement

Serge Wroclawski emacsen at gmail.com
Mon Jul 19 23:00:20 BST 2010


On Mon, Jul 19, 2010 at 5:17 PM, Sunburned Surveyor

> I was thinking mostly about the classification of road types. It seems
> that each country would have its own system for classifying types of
> roads. There is some documentation on the wiki about this. I suppose I
> was thinking OSM US might benefit from a documented road
> classification scheme. For example: If your road has 4 lanes, is
> paved, and is separated by a physical median your road type is
> "divided highway". Perhaps this type of system is already in place.

This has been discussd in the past. While there's certainly an
opportunity to import this data if it's already present in another
dataset that we have access to, import it- but I think  one of OSM's
best features is its ability to distill various local standards into
something that's meaningful and organic.

There are a lot of people who are interested in tag mapping (in the
sense of associating tags). There was a talk about it at SoTM. But
right now it's all mostly theoretical.

> Serge wrote: "This sounds like documentation questions. I don't know
> what a "route"
> is though, in OSM terminology. Do you mean a tagged road or do you
> mean something else?"
>
> I was thinking about acheiving more consistency again. If I have a
> long city street, is it represented as a single feature, or as several
> features connected at their end points?

It would be valid in OSM to respresent it either as a single way, or
as a series of ways in a relation. The tagging system abstracts that.

OSMers can go into long discussions on the right way to represent
something. The bottom line is basically that we can suggest "best
practices", but in the end there's a reality that people will do what
they want for the most part.

Efforts to automate cleanup of such things often result in disaster.

> If I do represent a long
> street as several features, where do I start and end them? If I
> represent a long street as a single feature, where do I put the nodes
> in the feature?

These are excellent questions for the new OSM help system help.openstreetmap.org

> Serge wrote: "So, yes to local groups. Yes to working with governments. Yes to
> working with universities, non-profits and buisness. No to formal
> structures and committees telling people what to map. That's just MHO."
>
> I think our e-mail exchange reveals a different way of looking at and
> thinking about OSM. Most people probably approach OSM as a fun hobby
> mapping project. I approach OSM from the perspective of a geospatial
> and mapping professional. I'm looking at OSM and wondering how I
> improve the consistency and accuracy of both the tags and the
> geometry. I suppose because I'm a surveyor I automatically start
> thinking about standards, guidelines, and committees. This isn't
> necessarily a good approach to OSM, but is the way I inherently try to
> tackle problems. I realize it is very non-organic.

OSM contains people from a variety of backgrounds, from hobbyists with
no geographic training to GIS professionals, to survoyors and members
of national GIS organiations.

Two large differences between other projects and OSM:

1) In other projects, the "what to map" is clearly defined. You have
road surveyors, or you have other organizations feeding data into a
GIS system.

In OSM, this is not the case. People map what they're interested in.
For example, I don't drive, so aside from what's in areal photography,
I don't map roads. I do map points of interest though, and
occasionally walking trails.

The strength of OSM is that its flexible enough to handle pretty much
any feature that's relatively stable and verifiable. So we have people
putting road data in, but we also have a kid in Germany whose been
mapping the utility polls and has an extremely detailed and
comprehensive map of the electrical system in his part of Germany,
including the number of power lines, the voltages, the companies who
own the lines, etc.

OSM is flexible and designed for either.

2) In most other maps, someone is paid to do the work.

There's the old saying "He who pays the piper calls the tune", well
what happens when there is no piper? That's why people get nervous
when they hear talk about imposing rigid standards on the map, or
creating comittes. People put up with that when they're paid, but I
think they wouldn't if they aren't.

Anyway, I think what the various Free/Open Source Software projects,
and Wikipedia have shown us, is that there's a great deal of power in
using as little structure as possible.

> I certainly don't want to kill anybody's OSM buzz. OSM is very
> organic, which is part of what makes it cool. It is also somewhat
> halphazard, at least from a traditional spatial data perspective.
>
> Ideally there will be a way to improve the quality and consistency of
> OSM data without killing the fun. It is obvious from my e-mails that I
> don't have the answers to that challenge. But I do think about the
> challenge and wonder how (and if) it can be overcome. Of course, this
> challenge isn't unique to OSM, and applies to all types of volunteer
> collected geospatial data. The challenge may be overcome through tech,
> and not through guidelines or standards.

I think the answer is "It depends on what the data is used for.".

There was a discussion at SoTM about city maps with MapOSMatic. One of
the audience members complained that with a city map, the data should
be in the proper local projection. That's a legitimate question, but
at the same time, for a simple one page street map, it won't make much
of a difference for simple navigation.

That's sort of how I see OSM. It's "good enough" and there's a lot of
value in "good enough", especially when our good enough becomes better
than someone else's perfect. :)

> Serge wrote: "My initial question to the list was "What can OSM US (the
> organization) do to involve the community more?"- I didn't really see
> that addressed in your email. Do you have any thoughts on it?"
>
> I didn't even know that the OSM US chapter existed. I just subscribed
> to the mailing list because I thought I might have questions about how
> OSM mapping was done in the US. Then I made the mistake of responding
> to that e-mail you posted. :]
>
> Perhaps it would be easier to engage the community if the mission and
> activities of OSM US organization was better marketed. If you guys
> aren't tackling standards and guidelines, what are you doing?

Our focus thusfar has been:

1) To create the scaffolding for the organization. We're currently run
by a temporary board that was put in place so that we could do the
work of founding the organiation (writing the bylaws, incorporating,
setting up membership, etc.).

2) To run a conference in the US. That event is SOTM US, which will be
happening in about four weeks in Atlanta, GA. I encourage you to
attend. http://sotm.us

3) To make it as easy as possible for the first elected board to run
the US organization easily.

Along with that, we're focusing energizing the US community.

> Are you
> developing mapping tools? Are you organizing campaigns to fill holes
> in the US OSM map? Are you showing people how to use OSM data to make
> beautiful maps or to answer real-life problems with a spatial
> component? Are you coordinating data collection efforts with other
> organizations in the US?

I think some of those are outside the scope of the US chapter. Others
aren't, but are difficult to accomplish at the moment with our limited
size and budget. We only started accepting members a two weeks ago!


> Serge wrote: "Your email had a lot of interesting ideas. What plans to do you
> have to follow through on them? And is there something you'd like to
> community to do to help you follow through in achieving some of the
> goals you mentioned?"
>
> Here is the list of things I'd like to accomplish to integrate my own
> local mapping work into OSM:
>
> (1) Create some guidelines for feature classification, tagging and
> topology that I discussed earlier. This won't be anything too formal,
> just something to guide my own work.
>
> (2) Figure out what metadata I want to track for the spatial data I
> want to put in OSM. Map this to a consistent set of metadata tags.

I'd suggest getting on the tagging list.

> (3) Set-up some sort of simple repository to release my spatial data
> under the public domain before upload to the main OSM database.

Imports are a touchy subject in the OSM and the US in particular.
There be dragons.

> (4) Create a set of open source tools to move OSM data in and out of
> OpenJUMP and AutoCAD.

Very cool.

> (5) Throw up a web page I can use to share these guidelines and tools
> with others.

Is there something you feel is missing in the OSM wiki and help site?

- Serge




More information about the Talk-us mailing list