[OSM-talk] Where came the concept of ways, segments, nodes?

Andy Robinson Andy_J_Robinson at blueyonder.co.uk
Wed Mar 21 11:14:35 GMT 2007

Mike Collinson wrote:
>Sent: 21 March 2007 1:34 AM
>To: 'OSM'
>Subject: [OSM-talk] Where came the concept of ways, segments, nodes?
>At 12:04 AM 21/03/2007, Raphaël Jacquot wrote in another thread:
>>the definition of a way is typically in the GIS world
>>* goes from one intersection to another intersection *
>>whoever says "the common definitions from the GIS world don't apply to
>>us because <blah>" is mistaken.
>Learning, ("Robbing" :-) ), from other disciplines is always a good thing
>to do.  Why reinvent the wheel?
>May I ask some of our more long standing members where the concept OSM
>ways, segments, nodes came from?
>Once I got over what I think is most newbies mental block about the
>different roles ways and segments play, I think, and still think,it a very
>elegant model, difficulties in rendering at the segment level
>notwithstanding.  I'd like to do some background research if there is
>anything published in the academic world.

This is from what I know and recall of the history.

SteveC is the oracle on the original database tables. It's worth though
noting the roadmap here, although I believe only he can give chapter and
verse on the original concept decisions.

Nodes and Segments were used from the outset of the database 2 years ago and
were certainly there and in use when the system went properly online, I
think that was around July 2005. At that point there was the concept of a
selection of segments being referenced together to define objects with
common information (eg streets). There was also an extra concept that would
cover areas. However neither of these two data models were user facing at
the time (although I think SteveC had some aspects of the tables set up)

At this point in time there was a little tagging being done using the
"class=" approach.

At the end of 2005 we started thinking more about tagging and grouping the
segments. Steve had been referring to them as Streets and Imi as tracks. I
came up with Ways because it seemed to fit better with the multi transport
model and I was at the time working on the initial version of Map Features,
which went onto the wiki in March 2006.

So, from the outset of ways and tagging there was never any thought that it
needed to be a rigid structure. That’s why ways could be any jumbled
container of segments and tags could be any key or value. The greater need
for a more logical and stricter way type only came about when 80n produced
the first version of Osmarender, also in March 2006.

Actually it's worth noting that both Osmarender and Map Features are exactly
a year old. We've done a lot with both in a year!

This all points to one thing, that the need for change has been diven by the
need to do something with the map data. We don't in my view need to change
something for the sake of changing it. We should only do so if to produce
the next cool thing makes a change necessary.

Hope I got the history lesson roughly right.



Andy Robinson
Andy_J_Robinson at blueyonder.co.uk

>talk mailing list
>talk at openstreetmap.org

E-mail message checked by PC Tools Spyware Doctor (
Database version: 5.06900

More information about the talk mailing list