[OSM-talk] JOSM and swedish characters

[Immanuel Scholz]

Hi,

> I am running Debian unstable.  So what is the methot to get 8-bit
> characters to work in JOSM?  I managed to input å,ä ens é and the look
OK in the property box.  But when I save the file (Files->Save) I get 
> 
>   Rydsvägen (Rydsvägen)
>   MÃ¥rdtorpsgatan (Mårdtorpsgatan)
> 
> My settings is
> 
>   rice:~> echo $LANG
>   C
>   rice:~> echo $LC_CTYPE 
>   sv_SE
> 
> and I do not use UTF stuff.

This is a known problem of the server. The bug is, that the server code
escapes UTF-8 byte values in html-encoding and send the the
html-sequences in ASCII-encoding instead of sending them right away as
UTF-8 bytes.


> A JOSM question.  How do you select the segments in a way?  And how do
> you choose between different ways containing the same segments?

To select segments under a way, hold down the Alt key when clicking.
(When you window manager capture Alt-Click, then press mouse, press Alt
and finally release the mouse button helps sometime ;).

To select between two ways, the simplest would be to click on an area,
where the ways don't overlap ;-)

If this is not possible, select both ways using a rectangle (Pressing
Alt in rectangle selection will select anything in touch with the
rectangle, not just everything completely covered). Then open the
selection dialog and double click on the way of choice.


> I must say that I am really impressed with JOSM.  I looked at it to
> get inspired for osmpedit but when I noticed that the user interface
> was of a type I like (minimal number of actions and no menu levels tha
> you have to use often) I did not help my inspiration for developing
> osmpedit...  Maybe I should learn more about databases and Ruby
> instead...  I have updated osmpedit so it supports ways somewhat and
> the perl classes are updated to handel ways.

Of course, as author of an offlice client, I am not in a position to
persuade you to abandon your editor. ;-)

But If I am honest, I don't think we need that many different offline
clients. And since there is plenty other things to do for a good perl
coder... I can't imagine we could ever run out of tasks.. ;-)

- There are tons of API tasks at trac, worth looking at. Many of them
require some improvements in the server deployment infrastructure, but
some can be done independently (15, 17, 54, 94, 120...). And ruby is
similar to perl in many ways ;)

- An offline cache for OSM data and background images would be cool.
Configuring squid isn't that easy and squid is suboptimal in caching
osm-data requests anyway (unless we find a way of normalize the requests
- this could be a task too ;).

- An proxy that alters the data passed by on a specific ruleset would be
cool. Tag replacement, reordering, filtering of data etc.. As example,
the proxy could filter the data and only pass major roads at lower zoom
levels. This would enable clients to view only relevant data. (This
could be combined with the offline cache?). Perl is maybe the best
language for extensive "regexping" out there ;-)

- Webservices that convert between data formats are nice. Convert from
OSM to GPX, MySql-dump, WMS (means:providing an WMS server), SVG, even
PNG would be cool. Converting from GPX to OSM is nice to have too.

...


Ciao, Imi.