[OSM-talk-be] 4th meeting STUK on 06 april 2011

[Ben Laenen]

Jo wrote:
> I was glad Jan-Willem showed up. The result is a python plugin which can
> correct spelling from within JOSM. I just corrected 264 street names around
> Aarschot with it:

Really great news, but I do want to warn about these automatic scripts. You 
can make mistakes very easily with them.


> > corrections = [('Dr. ', 'Docteur '),('R. ', 'Rue '),('Av. ', 'Avenue '),
> > ('Bd. ', 'Boulevard '),('Sq.', 'Square'), ('Pl.', 'Place'),

These are obviously meant for French places, but since I don't see some way to 
make the distinction: never run this script on a Dutch place. I also wonder 
how safe the "R." is in a French place, since it can easily be the 
abbreviation of a first name.

> >                (' De ', ' de ')

This is problematic. "De" can still be (and most of the times is) uppercase in 
Dutch. The only place to check this is the official street name list.


> >                , (' La ', ' la '), (' Du ', ' du '), ('Des', ' des '),
> >                (' Au ', ' au '),(' Aux ', ' aux '),('À', 'à'),

> >                (' Den ',' den '),

Same here.


> >                (" D'"," d'"), (" L'"," l'"),

> >                ("' ","'"),

Bad idea in Dutch.

> >                ("Ecole ","École "), (" Voor ", " voor "),

> >                (" Op ", " op "),

In Antwerp we have for example "Jos Op De Beecklaan": uppercase.

> >                (" Pour ", " pour ")]


You've probably thought about these things yourself already, but I did want to 
point it out before someone took the script and uses it. Also, always manually 
check the results from this script before uploading.

Greetings
Ben