[OSM-talk] Data primitives (was: The segments vs ways vs superways question again...)

[Richard Fairhurst]

David Sheldon <dave at earth.li> wrote:

> There needs to, and indeed can, be a distinction between the abstraction
> that users/editors use, the abstraction that programmers use and the way
> that the data is stored ultimately. Users shouldn't have to know how it
> is stored. At the moment there is only a very thin abstraction between
> the user and the database. We need to make this thicker rather than
> ruining the database.

Oh, I agree absolutely. That's why Potlatch hides segments from the  
user, offering only points and ways.

But you can't abstract everything away. The way that people input data  
will inevitably have an effect on the contents of the database. If you  
implement superways, but the editors set them purely programmatically  
(because, say, they're too complex to portray through the UI), then a  
superway becomes a predictable derivation from a set of ways and  
(IMHO) there's no point in having it.

To restate, we're more likely to get global coverage if we have a  
people-centred database*, rather than expecting people to behave in a  
database-centred fashion.

cheers
Richard

* That sounds horribly like Microsoft's current advertising campaign.  
I apologise.