Re: RFC: UTF-8 iCalendar bug solution

[Robert_Ransdell@xxxxxxxxxxxxxxxx]

The MIME header is the IMIP not the
ICAL.  If you save the mime stream it is not a .ics file;  It
is an .eml file.
_____________________
Note: new email address

tom_ransdell@xxxxxxxxxxxxxxxx

Doug Royer <Doug@xxxxxxxxx>
Sent by: owner-ietf-calendar@xxxxxxxxxxxx

02/26/2003 11:26 AM

Please respond to
"ietf-calendar@xxxxxxx" <ietf-calendar@xxxxxxx>

To	"ietf-calendar@xxxxxxx" <ietf-calendar@xxxxxxx>
cc
Subject	Re: RFC: UTF-8 iCalendar bug solution

John Stracke wrote:
> 
> Doug Royer wrote:
> 
>> >  I see popular iCalendar clients (with more
>>
>>> in the pipeline) using a web server to serve/transport iCalendar
>>> data. The web server will not (and can not reliably) transmit
the 
>>> character set encoding.
>>
>>
>>
>> Absolutely not true - HTTP transports MIME data and it DOES include
>> the MIME
>> headers including charset.
> 
> 
> The problem is that people expect to be able to take a text/calendar
and 
> save it to disk; when that happens, the MIME headers are lost, and
they 
> can't be reconstructed.  There's no way to look at the text/calendar
and 
> know what character set it's in; if there were, we wouldn't need the
> character set in the MIME header.
> 
> This is really a general MIME problem: MIME provides for transmitting
> metadata about objects, but that metadata tends to get lost once it
> leaves the MIME transport.  The only solution is to define 
> self-describing formats that don't need the MIME headers.
> 
> I hate to say it, but this is something that an XML encoding would
give 
> us for free.

No and for the same reason. If the XML file is in some uncommon
charset then you can not read it to find the encoding value.

                
<xml version="1.0" encoding="one-you-do-not-expect">

>>> 2. make a suggestion so we can all interoperate with the mixed
>>> character encodings until all vendors fix their software.
>>
>>
>> We did solve that problem :-)
>> That is why we mandate that iCalendar objects are MIME objects.
> 
> 
> Mandating it doesn't make it happen, unfortunately.

It does as far as the standard and interoperablility is concerned
and that is what the WG is about.

If a vendor wishes to save the data in the SHIFT_JISX0213 charset
they can. However sending that raw file to most implementations
will result in breakage.

> The only mandate that would work would be to declare that UTF-8 isn't
> the default; it's the only option.  People would gripe about
that (for 
> Europe, UTF-8 costs more than Latin-1; for Asia, it costs more than
> UTF-16), but it'd be an easy requirement to understand and implement.

And they did :-)
That is why the MIME header is not optional.



-- 

  Doug Royer                
    |   http://INET-Consulting.com
  -------------------------------|-----------------------------
  Doug@xxxxxxxxx              
  | Office: (208)612-INET
  http://Royer.com/People/Doug   |    Fax: (866)594-8574
                    
            |   Cell: (208)520-4044

                 We Do Standards
- You Need Standards