[Talk-us] OpenStreetMap Isn't All That Open, Let's Change That and Drop Share-Alike

[Greg Morgan]

In system management, success is reaching the next bottle neck.  My take
away from Alex's concern is "What is the next bottle neck for OSM to focus
on?"  We can already see some of the great successes that MapBox has
created from the ideas used to make a slippy map.  I am so thankful for the
whole idea of mbtiles.  What a wonderful idea.  I am so thankful for
TileMill.  I show people TillMill and how they can use it to get into GIS
with a limited budget. These people can save a ton of cash and learn the
same important concepts behind GIS without more expensive tools.  However,
I don't believe that the license is the next bottleneck.  The next
bottleneck to face is something more like creating the iD editor.

I've just described the gift culture of the open source/data world. My
dentist doesn't get it.  He thinks that I should make as much money as I
can.  Why are you giving this away?  MapBox has gifted us some great
things.  Part of my gift back is to use a project or tell other people
about these gifts.  Corporations don't always get this idea.  We can look
to the recent example of Oracle Corporation and the Jenkins/Hudson
controversy.  In short, Oracle wanted to take complete control of the open
source Hudson project.  The rift was so bad that Jenkins was forked out of
Hudson so that the project could continue.   In a perfect world, I'd say
Alex, let's change the license so that we can share things better with one
another.  However, I can't.  Others don't understand that gift culture like
you and I do.  From all bad things that I've seen, I wouldn't license the
data under any other license than ODbl.

ODbl isn't the only license that causes legal heartburn.  I've transcribed
part of Dr Richard Hipp's Google Tec talk,
http://www.youtube.com/watch?v=giAMt8Tj-84#t=1703, concerning his
experience with the public domain license:
"Observations from running the project...SQLite is in the Public Domain.
I've learned over these six years that this is not necessarily a good thing
from the point of the people who use it. I thought that when I was going to
release SQLite source code into the public domain that this would make it
easier for people to use it.  But big company engineers want to use it in a
product.  The engineers talk to the lawyers.  The lawyers see the public
domain and they panic. Oh what do we do?  We know how to deal with the GPL.
 We know how to deal with the MIT license or the Berkley license.  People
tell me that there is no public domain in Europe.  This is purely an
American concept.  I have generated an enormous amount of billing for
corporate lawyers by placing this in the public domain.  One company in
Europe was not willing to accept that this was in the public domain. So
they paid a small token fee to me and the other developers.  They made us
sign professional use licenses so that they would be covered.  So I don't
have an alternative what could be better. I think that this is a political
thing.  It surprised me how difficult it would be to use public domain
software."

Alex states a great ideal with, "Ideally the government of New York City
would just copy changes from OpenStreetMap to help maintain their own
datasets - but they can't. Many datasets managed by government behind
closed doors today should just be managed by the same maintainers on
OpenStreetMap tomorrow - with gains for everyone".  One of the problems
with the OSM database is that there is no guarantee of a permanent primary
key.  If I need to split a building, the original primary key does not stay
with one of the two pieces of the building.  Two new buildings are created
with two new primary keys.  I don't know of a way to bridge two systems
when you lose a primary key.  One way would be to store the primary key in
with the OSM data.  I believe this kind-of "meta" data is frowned on in the
community.  The meta data that would be required for joint maintenance can
always be deleted. A license change will not solve the government sharing
problem.  They have to find a way to jump in with both feet.

Next problem with this ideal is that OSM cannot guarantee the integrity of
the data once it has been placed into the OSM database.  I started to map
pedestrian crossings. I am OK that a crossing is not rendered at this time
because I use a node for a crossing, a node for a traffic light, and
another crossing node to help line up intersections.  I also thought that
it would be a great way to sell governments on the use of this kind-of data
to maintain their annual restriping of crosswalks.  A new mapper comes
along and starts deleting the crossing data because cartographers have not
rendered the data.  The data were/are considered map clutter.  I guess the
reason is that if I cannot see the data rendered, then why do I have to
plow through it in an editor?

Next problem with this ideal is tagging mistakes. The idea of iD editor was
to bring in new editors.  Invariably, mistakes are made.  I thanked a
mapper for his contributions in the metro Phoenix area.  Recently, I've
seen chunks of road disappear. Well come to find out that an extensive part
of his work was rolled back because he changed the highway value to
something else.  I thought the changes in the Phoenix area were OK and
welcomed his help in an actual thank you note.  The same view must not have
been seen the same in other areas of the US.  The missing data from the
rollback reminded me of what happened during the license change to ODbl.
 There were lot's of nodes no longer connected to ways.  The community
needs to be more welcoming in these cases than rolling back the data.  If
we want to encourage new mappers, then how to we do a better job guiding
the new mappers along?  A physical meetup model isn't always possible.  The
new mappers will more than likely be armchair mappers.  Licensing has
nothing to do with the new mapper problem.  Unwelcoming articles like the
Armchair Mapping page are not the solution either
http://wiki.openstreetmap.org/wiki/Armchair_mapping.  There are areas of
Phoenix I will not survey for safety reasons.

ARNOLD, http://www.fhwa.dot.gov/policyinformation/hpms/arnold.cfm, would be
a perfect use case for Alex' ideal. I had an interesting talk with an
individual involved with the State of Arizona's compliance.  License issues
were not even a concern.  The state decided to use all of the
municipalities 911 data for the road network over both TIGER and OSM data.
 The governor of each state has to sign off on the report. Imagine the
governor signing off on OSM data where data has been deleted because it is
map clutter or changes being rolled back because of highway values being
changed.  The state's project knew all about both TIGER and OSM data
availability.  The license question did not enter the discussion.  The
concern was what the governor would have to signon.

OSM is a unique project. OSM is even more unique that, say, Wikipedia that
also has a database.  In other projects, the gift culture of contributions
are split over forums, mailing lists, bug trackers, testing, build systems,
and the most important piece of the project is the code.  A "normal open
source" project  has no real easy way to aggregate all of these
contributions together like what can be done with OSM data.  The very
nature of a database login means that you can track it all.   Don't get me
wrong here but the OSM source code is not as important to the project as it
is to other projects.

My opinion is that many OSMers having been reading the OSM data wrong for
years.  All projects have attrition.  All the projects have lurkers.  In
the case of OSM, a lurker might be an OSM account that has not contributed
a change set to the database. Other projects struggle with losing mind
share to other projects or how to get the lurkers to contribute.  The
recent iD editor has been another attempt to get the OSM lurkers out of the
corners or attract new contributors.  Pascal Neis has a wonderful summary
of how well that solution has paid off by looking at each mapper's editor
use.  We can quantify lurkers and top contributors with the OSM database as
well as map the world.  I have no problem with the attrition rate or the
lurkers.  These attributes are part of an open source project. 30,000
active contributors is a great number.  As for lurkers in any project, it
takes guts to make the first public contribution.  Some contributors fear
making a mistake more than others.  In the case of OSM, my fear of making a
mistake on the data was irrespective of if the data was imported, armchair
mapped, GPSed, surveyed or what not.   I tell new mappers don't fear the
edit.  There's always OSM Inspector or KeepRight to help correct any
mistakes later.

And now for something completely different: It's the cartographers that are
limiting the project adoption and burnout rate for contributors.   Here's
your next bottle neck to work on.

I understand how difficult it is to render the world.  However, I consider
that the current method of data distribution and the rendering chain is
broken.  I don't have a better solution yet but here's a problem that is
right in MapBox's wheel house that can be fixed without a costly and
damaging license change. When Alex says, "OpenStreetMap's purpose is to
democratize who decides what's on the map" that's not true based on what is
actually rendered.  It is an oligarchy of cartographers that determine what
is on the map and not the democracy of what the mappers have placed in the
database.

Vector tiles are not the solution if the resulting tiles are the just more
of the same minimalist map tiles.  We need a real mapper's map again.  We
need tiles that are so butt ugly only a mother would hang the project tiles
on her fridge because that's what little Johnny did in school today.  The
type of butt ugly tiles I am talking about are something like Tiles at Home,
http://wiki.openstreetmap.org/wiki/Tiles@Home.  The magic of T at H was that I
was rewarded as a young mapper.  It was magic when I saw the first
traffic=hump that I added to the database show up on the map.  From there,
it was an Easter egg hunt to find traffic calming humps as I could while I
fixed tiger data.

The cartographers are creating beautiful maps. I show people MapBox maps,
Stamin Design, or Andy Allen's work.  However all these are just minimalist
maps that do not show what the database is capable of and the type of data
that the database contains.  Many people have to see it to believe it. Just
as the iD editor has been viewed as an important piece of the puzzle, I
view a mapper's map like Tiles at Home equally if not more important than the
iD editor. Here's an example. I mapped a ta ta bar,
http://www.openstreetmap.org/node/1801776750, It doesn't render. Another
mapper put a poi of a nightclub,
http://www.openstreetmap.org/node/2354317861.  It doesn't render.  That was
part of that mapper's first and last contribution.  The OSM process appears
broken.  In Google maps, however, a mapper using map maker is rewarded with
their efforts.  You can go to this same location and at least see what I
call a "Google Donut"
https://www.google.com/maps/place/The+Candy+Store/@33.6555233,-112.0307235,20z.
  The perception is that map maker works while iD, Potlatch, JOSM, Mercator
are all broken.  Moreover, the mapper's map tiles can't be some side
project that goes away like T at H did.  The butt ugly tiles need to be in the
second slot on the OSM site and managed by OSM servers.  In a sense, I have
off-boarded part of my mapping just like this mapper did.  I no longer map
data of this type.  Why should I even put a sport tag with a leisure=pitch
when all the the current tile set needs is is pitch to show a green blob?
T at H would reward me with an icon that showed the actual type of playing
field.  Being rewarded for mapping efforts encourages additional growth of
the database.   Does nearing completion of the map mean showing highways
only as all the current pretty map tiles show?

A license change is not what OSM needs.  The Linux project held firm to
their license even though businesses complained.  Corporations are
contributing to the project now.  Even Microsoft has contributed to the
kernel after years of calling the GPL a cancer.  OSM needs to hold firm to
the license that we have.  As people have pointed out a permissive license
would allow companies to just sweep the OSM data into their database
without gifting back.  I've spent a great deal of my resources in the way
of enjoyable time exploring and mapping my part of the world.  That has
been my gift to the project and fellow mappers.  Businesses need to figure
out how to join in and what they can regift to the project.  If they can't,
then there are always other paid alternatives to OSM data.  They have to
weight the perceived cost of giving up data verses the cost of paying a
service provider like Google to keep their IP.

I hope this helps,
Greg


On Thu, Mar 13, 2014 at 7:26 AM, Alex Barth <alex at mapbox.com> wrote:

> Hello everyone -
>
> I've been sitting on writing about the detrimental effects of
> OpenStreetMap's share-alike license (ODbL) for a while and finally decided
> to, um, share. I've been listening long to many OpenStreetMappers I respect
> a ton telling me it's not so bad and it's just what we're stuck with right
> now. But given how bad share alike is for OpenStreetMap I don't think we
> should give up for pushing for a more open license. Here's why I think
> share-alike hurts OpenStreetMap and how this keeps OpenStreetMap from
> having the full impact it could have:
>
> http://www.openstreetmap.org/user/lxbarth/diary/21221
>
> Looking forward to your comments,
>
> Alex
>
>
> _______________________________________________
> Talk-us mailing list
> Talk-us at openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk-us
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/talk-us/attachments/20140316/3331180c/attachment-0001.html>