[OSM-dev] Cloudmade routing for OSM rails_port site.

Nick Black nickblack1 at gmail.com
Thu Apr 30 17:42:48 BST 2009


On Thu, Apr 30, 2009 at 1:56 PM, Andy Robinson (blackadder-lists) <
ajrlists at googlemail.com> wrote:

> Nick Black [mailto:nickblack1 at gmail.com] wrote:
> >Sent: 30 April 2009 1:48 PM
> >To: Andy Robinson (blackadder-lists)
> >Cc: Frederik Ramm; dev at openstreetmap.org; Chris-Hein Lunkhusen; Tom
> Hughes
> >Subject: Re: [OSM-dev] Cloudmade routing for OSM rails_port site.
> >
> >
> >
> >On Thu, Apr 30, 2009 at 11:44 AM, Andy Robinson (blackadder-lists)
> ><ajrlists at googlemail.com> wrote:
> >
> >
> >       Nick Black wrote:
> >       >Sent: 30 April 2009 9:04 AM
> >       >To: Frederik Ramm
> >       >Cc: dev at openstreetmap.org; Chris-Hein Lunkhusen; Tom Hughes
> >       >Subject: Re: [OSM-dev] Cloudmade routing for OSM rails_port site.
> >
> >       >
> >       >
> >       >
> >       >On Wed, Apr 29, 2009 at 9:45 PM, Frederik Ramm
> <frederik at remote.org>
> >wrote:
> >       >
> >       >
> >       >       Hi,
> >       >
> >       >
> >       >       Dermot McNally wrote:
> >       >
> >       >
> >       >               Clearly every potential user of OSM is different,
> but
> >for
> >       me, a
> >       >key
> >       >               benefit when I discovered the project what "hey,
> >vector
> >       data,
> >       >that can
> >       >               be used for routing!". If we think that others
> >discovering
> >       the
> >       >project
> >       >               would thing likewise then a prominent path from
> >project home
> >       >page to a
> >       >               routing engine (and I really couldn't care less
> >whose) would
> >       be
> >       >a Good
> >       >               Thing. As long as it works acceptably well, of
> >course...
> >       >
> >       >
> >       >
> >       >       But isn't this angle one where the open/closed distinction
> >gains
> >       >weight again? Currently, you can (and I've done this a number of
> >times)
> >       >tell potential OSM users: "Everything on openstreetmap.org is
> free
> >software
> >       >and free data - you can build the exact same site for yourself
> with
> >all its
> >       >functionality from our database and our svn" *). With non-free
> >routing or
> >       >geocoding services that would stop.
> >       >
> >       >
> >       >If we want OSM to grow and become more successful than it is then
> we
> >have
> >       >to get over the idea that the only people who matter are software
> >       >developers.  There are around 12 - 17 million software developers
> in
> >the
> >       >world.  How many of those developers know how to use SVN?  Maybe
> 1%.
> >How
> >       >many are FLOSS developers?  How many care enough about OSM to
> learn
> >how to
> >       >use it?  If 1% of software developers care about OSM, which is
> very
> >       >optimistic estimate, then we have a total pool of people who could
> >do as
> >       >you suggest of around 170,000 people.  Not very many.
> >       >
> >       >How do people discover OSM?  Their friend tells them or they find
> it
> >       >through a Google search.  They take a look at the front page of
> the
> >site
> >       >and within 2-5 seconds they make their minds up about it.  I don't
> >have the
> >       >numbers, but I'd guess that we have between a less than 1%
> >conversion rate
> >       >from unique visitors to OSM to people who sign-up for an account.
> >Imagine
> >       >if we could increase that rate by 5x.  5x more people sign up for
> an
> >       >account, 5x more people mapping, 5x more people submitting patches
> >to JOSM
> >       >and Potlatch.  5x more people at SOTM, 5x more Foundation members.
> >       >
> >       >We really need to think about what is more important to OSM.  The
> >most
> >       >important thing in my view is creating the best free map of  the
> >world.  To
> >       >do that we need to encourage more people to join OSM - people who
> do
> >not
> >       >know or care what SVN is or why a routing algorithm does its
> thing.
> >       >
> >       >Surely the biggest victory for any open source project is to
> squash
> >its
> >       >competitors and completely change the game?  Few open source
> >projects have
> >       >achieved this.  In 5 years time, the entire concept of proprietary
> >maps
> >       >created by guys in vans could be a thing of the past.
> OpenStreetMap
> >and
> >       >our way of doing things will have completely wiped the floor with
> >our
> >       >competitors.  We can only get to this point if we stop thinking of
> >       >ourselves as a small, specialist community and start to think of
> >ourselves
> >       >as a mainstream movement that caters for the needs of everyone -
> not
> >just
> >       >the free software world.
> >       >
> >
> >
> >       I feel this somewhat misses the point. Openstreetmap is about the
> >data and
> >       nothing else. No data, no interest in OSM, huge data, huge interest
> >in OSM.
> >       That interest will come automatically if we have the best data set
> >bar none.
> >
> >
> >OK, but there's a massive chicken and a bigger egg here.  The point is
> very
> >much that OSM should be doing everything we can to encourage a broad range
> >of people to contribute to the map and be part of the community.  What
> >really worries me is an attitude that the only people who deserve the more
> >advanced map editing tools that sit in OSM's SVN are people who have the
> >technical ability to use them.  I don't think any of us want to create a
> >project that is only accessible by the technical elite.
> >
>
> Yes, and I'll jump with joy if someone creates a very straightforward
> editor
> that just enables a person to add a POI or tag existing objects. We could
> encourage everyone from school children to pensioners in the local library
> to add stuff if we had that kind of tool in the toolset. A new routing
> service may excite a few but is not going to get the world adding data.


But its the first step towards improving the quality of data we have right
now and making it suitable for routing.  This is how I look at it:

* First there was no map at all. just an API.  Data was nodes and segments,
maybe some names
* Next was green landsat and white lines.  So people put in more lines and
the number of roads grew, we completed the M25
* Then 80n invented Osmarender and everything picked up.  Now there was a
point to tagging.
* Then we discovered Artem and Mapnik - and everything exploded because now
there was a point in tagging everything as you (and everyone else) could see
the results of your efforts
* Next up came no-names - now you are shamed into tagging the roads in
your neighborhood.  We thought London was nearly complete, but actually it
was far from it.
* Next up, routing...

The thing that each of these steps forward have in common is that they got
mass adoption.  Mapnik and the slippy map had such a bigger impact on the
quality of data in OSM than Osmarender because every one of the 1000s of
people who visited the front page could see it.  You didn't have to know
what Xalan was to have a map.  So the map got better.

I see the same thing with routing.  Right now, routing is restricted to a
segment of the community who can access it.  The problems with the data that
restrict it being widely used for navigation are just not exposed to most of
the community.

So the idea of getting routing into the forefront of OSM isn't about
transforming osm.org into a portal or OpenStreetMap or the Foundation
controlling the delivery of data - its about OSM doing everything possible
to make the best map possible.

The main way that thing get better in OSM is by exposing the problems -
thats exactly what would happen if routing was exposed to the mainstream
community, rather than the small number of people who hang out on mailing
lists, IRC and have the ability / time / geographic location to use the
current routing services.



>
>
> Potential new contributors won't be bothered whether the editor they use is
> open source or closed source (same as a GPS receiver now). As long as it
> sends data to OSM without any caveats then jobs a good un. But this in my
> view is very different from offering clever new services to the wider
> public. As others have said, where would that end? If everyone wants to
> change what OSM is or should be in the future that’s fine, but we need that
> debate first before we contemplate embracing a wider selection of consumer
> side services.
>
> Cheers
>
> Andy
>
> >
> >
> >
> >       From that point of view it doesn’t matter at all what software is
> >used to
> >       make use of the data, but as an open source project we should at
> >least
> >       encourage and promote any open source software that is found to be
> >useful
> >       and relevant. On the other side of the equation I don’t see it as
> >OSM's
> >       place to go promoting commercial products, even if they are good.
> We
> >can
> >       list them on the wiki etc, but surely it is for those commercial
> >companies
> >       to do their own promotion of OSM related products, not OSM.
> >
> >       As for thinking we are a small specialised community I stopped
> >thinking that
> >       way many months ago. We are THE big player here, but is in data,
> not
> >       software. OSM is not a software development project. Too many eggs
> in
> >the
> >       basket if we were. Lets stick to what we know, which is collecting
> >and
> >       maintaining a great dataset.
> >
> >       Cheers
> >
> >       Andy
> >
> >
> >       >
> >       >
> >       >
> >       >
> >       >       Bye
> >       >       Frederik
> >       >
> >       >       *) And add in fine print "if you're willing to devote a few
> >man-
> >       >months to get to grips with the finer details of Ruby, PHP,
> Python,
> >       >PostGIS, Mapnik and others"
> >       >
> >       >       --
> >       >       Frederik Ramm  ##  eMail frederik at remote.org  ##
>  N49°00'09"
> >       >E008°23'33"
> >       >
> >       >
> >       >
> >       >
> >       >
> >       >--
> >       >--
> >       >Nick Black
> >       >twitter.com/nick_b
> >
> >
> >
> >
> >
> >
> >
> >--
> >--
> >Nick Black
> >twitter.com/nick_b
>
>
>


-- 
-- 
Nick Black
twitter.com/nick_b
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/dev/attachments/20090430/ad26050f/attachment.html>


More information about the dev mailing list