[OSM-talk] Historical Data in OSM database

Frederik Ramm frederik at remote.org
Tue Nov 9 10:42:40 GMT 2010


Richard,

On 11/09/2010 01:23 AM, Richard Palmer wrote:
> 	I can continue to do this using a standalone copy of the OSM database
> 	but would prefer it to be made available for other people to
> 	improve and add to.

There are three reasons not to do this:

1. Mappers are not used to working with data like that - if they see 
that a shop is gone they will delete the shop, rather than researching 
when it closed and adding an appropriate end_date.

2. Our software cannot handle such data very well at the moment; ideally 
you'd wish for editors and/or the API to filter out irrelevant stuff so 
that someone mapping a present-day is not fazed by ancient Roman roads. 
The stat_date/end_date or other lifecycle concepts (search the Wiki for 
that word) are also not sufficient to describe what historians will want 
- they cannot for example model something moving from one place to 
another through history.

3. Our infrastucture might be unduly burdened by all the extra load.

All these things could be fixed with time if there was sufficient 
interest but it will require a lot of work and perseverance.

For now, I suggest that you keep your separate database. You can link 
your objects to OSM objects by their ID if you want, and process the 
daily change files to be alerted of changes in the objects of interest 
to you.

Bye
Frederik



More information about the talk mailing list