RE: Character encoding in xAP messages

["Edward Pearson" <edward.mailgroup@xxxxxxxxxxxxxxxx>]

------=_NextPart_000_0006_01C501EF.157C5B10
Content-Type: text/plain;
charset="utf-8"
Content-Transfer-Encoding: quoted-printable

My vote would be for the UTF-8 since I have encountered all the same
proble=
ms that Malcolm sees.

_____=20=20

From: Malcolm Green [mailto:groups@xxxxxxx]=20
Sent: 23 January 2005 23:19
To: xAP_developer@xxxxxxx
Subject: RE: [xAP_developer] Character encoding in xAP messages


If the data you're sending is 7-bit ASCII, then encoding using ASCII or
UTF=
-8 will yield identical results.  So an existing application which is only
=
using 7-bit ASCII should still work OK if you change to UTF-8.  As with
you=
r Exstreamer client, if you try to send data which includes an extended
cha=
racter set, it currently won't work correctly with ASCII encoding, but
will=
if you change to UTF-8.  So I don't see any downside to the proposed modif=
ication.  Incidentally, wherever filename strings are being sent via xAP,
I=
think it really ought to cater for Unicode strings as these have been supp=
orted in Windows filenames since NT4.=20=20
=20
I intend to release another version of TelCanto with further xAP support
th=
is week.  My current choices are:
- stick with the existing ASCII encoding, which fails to handle several
fil=
es in my music library correctly.
- change to using binary encoding, which will require a change to the
Audio=
schema and changes to all other applications which use it.
- use UTF-8 encoding which works correctly with both TelCanto and xAP
Deskt=
op as far as I can tell (I've not tried any other applications).
=20
I don't think the first option is viable in the long term, as it just
doesn=
't work correctly!  I obviously favour the third option, but am happy to
ta=
ke advice from those with a deeper knowledge of xAP than myself.
=20
Malcolm


_____=20=20

From: Edward Pearson [mailto:edward.mailgroup@xxxxxxx]=20
Sent: 22 January 2005 13:46
To: xAP_developer@xxxxxxx
Subject: RE: [xAP_developer] Character encoding in xAP messages


The spec says ASCII and, thinking it through, it probably means ASCII and
n=
ot UTF8.
=20
I'd previously observed accented characters not transferring correctly in
s=
ong titles between my custom Exstreamer app and its xFx controller. The
'fi=
x' to xFx below would allow these characters to be encoded but in a way
tha=
t's off spec. The Barix Control Language (tinyBasic) that the Exstreamer
co=
de is written in expects ASCII of course so (not tried it) will barf if it
=
starts to see double-octet characters in messages.
=20
Sounds like using the binary operator is the only safe way to go for
unicod=
e strings then. But I bet many xAP progs don't deal with this correctly.

_____=20=20

From: Mark Harrison (Groups) [mailto:mph@xxxxxxx]=20
Sent: 19 January 2005 23:24
To: xAP_developer@xxxxxxx
Subject: Re: [xAP_developer] Character encoding in xAP messages


Edward Pearson wrote:=20

Oh cool - you fixed an old bug in my Extreamer client that'd I'd never had
=
the time to track down!
=20
To formalise this - the xAP specification says ASCII everywhere. Was this
d=
one for a specific reason or were the specification definers just being a
b=
it loose here? Maybe the spec should be revised to say UTF8.
=20
Right, I'm off to listen to some German techno with umlauts in the title
an=
d other diacritically challenged corners of my music collection...

I had understood that non-ASCII stuff should all be represented in a
binary=
format, and therefore use the "keyword!" rather than the
"keyword=3D" form=
... and that the UTF-8 encoding should be specified as the encoding method
=
in notes section of the Schema definition

>From the point of view of my own code, I honestly don't know how, say,
xAPl=
ogger would respond if you tried to log a xAP message containing "Pot
s=C4=
=83 m=C4=83n=C3=A2nc sticl=C4=83 =C8=99i ea nu m=C4=83
r=C4=83ne=C8=99te." =
(which is, in case you were wondering, the Romanian for "I can eat
glass.")=
=20

Anyone got access to a suitable installation this evening to try?

I have no particular objection to a Specification (xAP 1.2.A ?) revision
to=
say "UTF-8" in appropriate places.

I seem to remember that ASCII was picked support by LCD / VFD display
devic=
es at a time when it was felt that these would be a common part of most
xAP=
installations.... I'm not sure how many have been produced, and whether th=
ey all support non-ASCII.

M.


_____=20=20