Now, Mark Keasling wrote, "If SHIFT_JIS, EUC-JP and ISO-2022-JP aren't supported you'd have a difficult time entering the Japanese market." But to what extent is such support going to be expected (or demanded)?
1) You will need to be able to send messages in ISO-2022-JP.
2) You will need to be able to read files on UNIX in ISO-2022-JP and EUC-JP.
3) You will need to be able to write files on UNIX in ISO-2022-JP and EUC-JP.
4) You will need to be able to read files on Windows in Shift-JIS.
5) You will need to be able to write files on Windows in Shift-JIS.
You've been able to use c-client to do (2) and (4) and convert into UTF-8 for a while now. c-client in imap-2004 has a new utf8_cstext() routine to do (1), (3), and (5) when the source is UTF-8.
But would someone developing a Japanese-market email client be content to work internally using only Unicode or would he want to actually handle the data in those other charsets?
It depends upon what you mean by "handle". You certainly must be able to write local files in EUC-JP and Shift-JIS (and probably also ISO-2022-JP), you must be able to send in ISO-2022-JP. And you should be able to accept input in "all of the above".
Once you do that, what you use internally is your application's own business.
We have two apparent options: (1) converting every text part to UTF-8 while keeping a record of the original charset or (2) giving the developer an option of reading text in either UTF-8 or the original charset. We would prefer (1) but are quite ready to provide (2) if necessary. And is there a (3)?
(3) is "both of the above".
I assume that for outgoing messages, a parallel situation would apply. Are there in fact any differences there? We can either (1) require that the application supply all outgoing text in UTF-8 and then convert it to any requested charset before the message is sent or (2) accept outgoing text either in UTF-8 and convert it as requested, or in any other labeled charset in which case we treat it as binary data.
For sending Japanese email, use either ISO-2022-JP or UTF-8. ISO-2022-JP is preferable; although many Japanese can now read UTF-8 email a sizable minority still can not do so. Only broken clients send email in EUC-JP or Shift-JIS.
-- Mark --
http://staff.washington.edu/mrc Science does not emerge from voting, party politics, or public debate. Si vis pacem, para bellum.
