[Openstreetmap] OSM 8bit database transparency.

[Tom Carden]

On 1/29/06, Johnny Doe <uucp1 at yahoo.com> wrote:
>
> --- Erik Johansson <erjohan at gmail.com> wrote:
> >
> > 1. You should use UTF-8, because that is what will
> > be used.
>
> utf8 is not a magic bullet, because it can utilize
> exactly
> these codes.
>

Although we've not been very mindful of character sets and so on, it's
unlikely we would ever use anything but UTF-8 without a very good
reason, so Erik's answer is ultimately correct: put anything you like
in the database, so long as it's valid UTF-8.  If it doesn't work,
file a bug in Trac.

> > 2. I have no idea what those characters look like
>
> There are no "characters" at these positions, but the
> codes can show up in the utf8 datastream.

Can you explain what they do?  Are these codes ones that you would
actually use, or are you just wondering?

> > 3. At this moment, it accepts what ever you put in
> > there.
> >
> If i will send a 'name=some0x80name' into the
> database,
> will i get back the same name?

If it's valid UTF-8, and we have no bugs, then yes.  Have you tried?

> The applet is not
> supporting
> utf8 right now, so its behaviour is irrelevant.
>

So certain it's the applet and not something else?  Perhaps the font
we're using might not support your characters, for example.  Can you
provide a test case?  There is a ticket about this already - one for
the applet, and one for the database.

http://www.openstreetmap.org/trac/ticket/28
http://www.openstreetmap.org/trac/ticket/29

If you have found things that do not work, then let us know there.

Tom.