Re: [cgmo-webcgm] Fwd: WebCGM 2.0 and ISO/IEC 2022 switching

[Lofton Henderson <lofton@rockynet.com>]

Here's my take on the question, embedded below.  I'd like to hear
from other TC members, if you have an opinion (and be sure to copy
Jeff).

At 09:45 AM 3/25/2006 -0700, Lofton Henderson wrote:

> > Date:
> > Fri, 24 Mar 2006 15:23:45 -0800
> > From: Jeff Wolkenhauer <jeffrey.wolkenhauer@ugs.com>
> >
> > =======================================
> > Hello group,
> >
> > I am hoping I can get some clarification regarding the use of ISO/IEC
> > 2022 switching within non-graphical strings in WebCGM 2.0.
> >
> > First, my understanding from the PPF is that
> >   1) ISO/IEC 2022 switching is allowed within SF strings.
> >   2) The only allowed encodings are ISO Latin1, UTF-8, and
> > UTF-16.
> > So a CGM generator may switch between any of those three
> > encodings.
> >
> > Is this correct?
> >
> > I ask because the phrase in the section T.14.5 of the PPF:
> >   "Otherwise, the use of ISO 2022 switching
> >    is prohibited in non-graphical text string."
> > confused me at first, but it's placement in the text and the other
> > references to allowing 2022 switching imply the meaning of the phrase to
> > be, "Use of character sets other than ISO Latin1, UTF-8, and UTF-16,
> > are prohibited for ISO 2022 switching in non-graphical text
> > strings".

T.14.5 is here:
http://docs.oasis-open.org/webcgm/v2.0/WebCGM20-Profile.html#webcgm_4_3

Quoting the text:
[[[

 The permitted character sets for non-graphical text are ISO Latin 1 (LHS & RHS), and UNICODE UTF-8, and Unicode UTF-16. Only one of these three shall be used throughout any particular WebCGM metafile instance. According to the CGM standard, the default SF character set, at the beginning of the 'metafile id' parameter of the BEGIN METAFILE element is ISO Latin 1. If the metafile is to use UTF-8 for SF parameters, then the following 4-octet ISO 2022 sequence shall occur as the first 4 octets of the 'metafile id' parameter: ...etc... 

]]]

I read this to mean that ISO2022 switching can *only* happen once, at the
beginning of the Begin Metafile id string, and that selected SF character
set then must pertain for the whole metafile.  (This is the way I
have read WebCGM 1.0 as well.


> > If
> > the answer to my first question is "yes", my second question
> > relates to the scope of the 2022 switch.  Is the character set
> > invoked only for that string (my hope) or for all subsequent SF strings
> > in the rest of the metafile?

I believe that the SF character set is established for all SF strings in
the metafile.


> > For
> > example, if the default ISO Latin is in effect, and a screen tip is to be
> > output in Japanese using UTF-8, I would prepend <ESC 2/5 2/15 4/9>
> > to the string.  But I am hoping that subsequent SF strings can be
> > specified in Latin again - especially since UTF-8 has no
> > return.

I see your use case, but in my view, it is not supported in 2.0, nor in
1.0.

Having said that, I do see one small loophole in the wording of
T.14.5.  It says that only one of the three sets is permissible
within a metafile.  As worded, it might be arguable that it in fact
would not prohibit the repeating of the ISO2022 string **for that set**
at other places in the metafile, in addition to the beginning of the id
string of BegMet.  But that would be a fairly useless thing to do --
re-declare the same set that's already in use.

It is my belief that CGMO intended that the only occurrence of ISO2022
switching would be at the very beginning of the metafile, at the start of
the BegMet string.

Jeff, if you are unable to use the "Join this TC" button on the
TC home page, I'll forward any reply of yours to the list, until we sort
it out.

Regards,
-Lofton.


> > In
> > clear text:
> > BegAPS "Eng_Ja_Howdy" "grobject" StList;
> > APSAttr 'name' "14 1 'Japanese Greetings'";
> > APSAttr "screentip" "14 1 '<ESC 2/5 2/15
> > 4/9>...UTF-8 for JP-Greetings'"; %
> > (ö %
> > APSAttr "region" "11 1
> > 1            %--
> > typ:idx, cnt:1, val:1(rect)   --%
> >                  
> > 16
> > 4             
> > %-- typ:vdc, cnt:4, val:(4-pts)   --%
> >                  
> > 520 859 708 914"; %-- Points defining a
> > rect        --%
> > BegAPSBody;
> >     RestrText 187 55 (520 859) FINAL
> >       "Hi";
> > EndAPS;
> >
> > So APSAttr _type_ "screentip" is ISO Latin, the screentip
> > itself is UTF-8, and the following APSAttr type "region" is ISO
> > Latin.
> >
> > Thanks very much for your help with these two questions.
> >
> > Best Regards,
> > Jeff Wolkenhauer
> >
> > UGS The PLM
> > Company       
> > (503.466.9401)
> >                                       
> > jeffrey.wolkenhauer@ugs.com
> >                                       
> > 15455 NW Greenbrier Pkwy, Suite 210 * Beaverton, OR 97006
> > --