[OSM-talk] Changing capitalization (Lima)

Phil Endecott spam_from_osm_talk at chezphil.org
Fri Jun 1 12:13:08 BST 2012


Steve Doerr wrote:
> On 31/05/2012 20:13, Worst Fixer wrote:
>
>> If you notice some big flaw in my case change algoritm, mail me privately.
>
> I'd only comment that a really intelligent de-capitalization algorithm 
> would attempt to supply the accents that are missing from the 
> capitalized forms.

I had to do this for a (non-OSM) Canadian dataset where the English
placenames were in mixed-case but the French (Quebecois) were in ALL
CAPS - presumably done deliberately to avoid the issue of accents.

The best approach is probably to use a dictionary generated from other
placenames.  In the case of OSM we already have a great source of such data.

I could probably do this for you, or at least provide a mapping table
that you could combine with your current conversion script.  But not
until next week.  Let me know if you're interested.


Regards,  Phil.










More information about the talk mailing list