Member Discussions

I have mudded for about 5 years and I am currently studying abroad in Japan. As I was thinking about new ways to use and practice this quite difficult language, I starting thinking about a MUD that was in Japanese. Now granted, there probably isn't a great demand for a Japanese MUD. The Japanese themselves mostly use cellphones for communication, most families outside of major cities don't have broadband internet access, and telephone connectivity is expensive. However, would writing a Japanese mud be feasible(sp?). The API's to send the data Unicode are in place, but I am unaware of a protocol that is widely available that will accept Unicode characters. Not to mention a new codebase would have to be written/adapted to use Unicode/Multibyte Characters, etc.

Any ideas are welcome :D

(Comment added by skyfw on Fri Nov 7 10:21:28 2003)

Edit: I guess a lot of the Protocol issues could be avoided by writing a cross-platform custom client that would produce the correct kanji. However, I would like to avoid this.

First off, a new codebase is not a necessity, but you certainly have your work cut out for you so a new one might be just as easy. C++ supports Unicode by w(ide-character)strings and locales. As far as clients go, MS Telnet should be adequate, if you take a look at this.
As far as the server side, here is a solution for a Unicode-enabled Telnet driver using MFC (look down on the page), and here is a (partial) solution for Unices.

Hopefully that will get you started. My codebase will hopefully include locales, but that's some ways off still :)
.

You might take a gander at this message. Seems related.

I'm currently a High School student that has several years of Japanese study. I'd love to see this. If you need any help, I am working on a MUD (english, though), I would readily offer support. Give me a line at my email.

C++ supports Unicode by w(ide-character)strings

Wrong. C++'s (and C's) wchar_t type is to hold multi-byte characters. Unicode falls into this category, but it isn't a requirement that an implementation must support Unicode through the wchar_t type. Also, many implementations have problems with multi-byte characters, so you need to be careful when using them.

The intent of locales isn't to support Unicode, too.

The custom client is the best solution for the maximum number of users to play on your potential MUD. The Telnet RFC requires only that a Telnet implementation should communicate in a 7-bit ASCII stream. While it doesn't limit implementations to only do that, I don't find a lot of implementations that do so otherwise. I've heard speculation that you could go ahead and use Unicode streams and it will work for Window clients, but I would think that limits the players to only Windows NT derived clients (since Windows 95/98/Me/XP Home? don't use Unicode).

There are Unicode libraries for all C++ platforms (bar maybe some embedded ones) and all these can be replaced so that the multibyte characters use Unicode -therefore, locales and wchars facilitate Unicode. I was admittedly unclear.

.

> The Telnet RFC requires only that a Telnet implementation
> should communicate in a 7-bit ASCII stream. While it doesn't
> limit implementations to only do that, I don't find a lot of
> implementations that do so otherwise.

It might be worth having a look at the Extended Ascii Telnet Option (RFC 698) and possibly the Character Set Telnet Option (RFC 2066).

There may yet be a Telnet client which does all the stuff you want it to do.

As a heads-up, I believe that TeraTerm Pro (http://hp.vector.co.jp/authors/VA002416/) will handle Japanese characters over a telnet connection. I'm not sure of the encoding they use, and my ability to read Japanese is about a word a minute, so I can't be much more helpful than that.

Last I checked, telnet only restricts the use of character 255 (which it uses for control codes), and in fact requires that lines be 8-bit-clean as part of the protocol. It is not restricted to 7-bit ASCII, although the results are a bit unpredictable if you use higher-order stuff (for example, on my machine, I tend to get Cyrillic characters instead of accent Western European ones, because of the code page I use).