From: The Radio Prague Staff of Highly Skilled Experts <barry@RADIO.CZ>
Subject: Re: =?ISO-8859-2?Q?=C8e=B9tina=3F_Pros=EDm!!!?=
Date: Fri, 22 Dec 1995 12:31:54 +0100
Next Article (by Date): Re: Cestina??? Prosim, ne! Milan Zamazal
Previous Article (by Date): Re: Cestina??? Prosim, ne! Matus Uhlar
Articles sorted by: [Date] [Subject]
On 23 Dec 1995, Stanislav Koci wrote: > >Takze myslim, ze cestina by mela byt, jsme =3DC8E=3D8AI nebo ne? :-))))) > > Coze jsme??? :-) (Excuse my English) What we have here are several problems. First, while there is a standards-track specification for support of many different languages (see the MIME RFCs) which should be a standard very soon, not everybody supports MIME yet. It is estimated that more than 50% of e-mail users have MIME support through the Internet, but sadly, it cannot be found everywhere. What I would hope is that the percentage of users with MIME support in areas where a character set other than US-ASCII is in wide use, would be much higher, such as closer to 90%. I believe that in countries such as Germany, where the push to be able to use the national characters is much stronger, the penetration of MIME-aware mail clients is close to this 90% figure, or even higher. I can see that here it is less, but it is not difficult to increase this. MIME-aware mail and news readers are available for nearly every platform. Often, MIME support can be added to existing programs that users are familiar with, such as ELM, or Emacs RMAIL, through simple patches or through the metamail program. Sometimes, to get MIME support, a user will have to move to a different mailer, and not all users are happy with that. The FAQ which is posted to comp.mail.mime and which should be archived at an FTP site near you tells more about programs which are MIME-aware. New programs which understand MIME are always being released, so the number of MIME-aware users is always increasing. Unfortunately, in CZ/SK, there is not a single standard for encoding of national characters, even on the same platform. This has been solved in most of western Europe by dropping support in mail for all the national code pages, so that nearly every mail reader supports ISO 8859-1 on the wire for sending and reading mail, translating to the local encoding. It is suggested that MIME mail in CZ/SK make use of ISO-8859-2 for widest interoperability. But since PCs and Windows and Macs use a character encoding which is different from ISO-8859-2 (only Unix machines are likely to have this as the native charset encoding), the mail program must do the translation. Unfortunately, not all programs do this. For example, I know that Eudora for the Mac can translate ISO-8859-1 to the Mac charset, but it leaves ISO-8859-2 alone. And it seems that other versions of programs, like Pegasus mail, also do not do the translation. > >Takze myslim, ze cestina by mela byt, jsme =3DC8E=3D8AI nebo ne? :-))))) ^^^ This message was first sent from Pegasus mail. In the copy I read from a news spool, I do not see any MIME headers. But I know that it is not in the ISO-8859-2 standard, first because I did not see the correct characters on my display with the original message, and second, because I know that a character which shows up with the QUOTED-PRINTABLE encoding (which is what most non-MIME-aware recipients will see) as ``=3D8A'' is not a valid character in ISO-8859-2 (it would be a control character). ``=3DC8'' is uppercase C with caron (hacek). So this is a problem that Windows has a few characters in the wrong location, and a MIME-aware program should do the translation from the code page to the desired ISO-8859-2. (MIME permits you to send the message as long as you tag it in the headers as CP-1250, but this would not show up properly with most mailreaders, so better is to use ISO-8859-2.) However, not all sites which are not aware of MIME QUOTED-PRINTABLE need to suffer, since it is possible to send data in mail as 8-bit characters, provided that you conform to the ESMTP specs for 8BITMIME. This requires a sendmail program that knows how to handle 8BITMIME, and the newest BSD Sendmail (8.7.*, presently 8.7.3) can do this. You also need to have a user mail program which can send 8BITMIME. Many mailers are presently set up to automatically QUOTED-PRINTABLE-ize everything, but there are plans for some which can use whatever is supported. I can speak for the next release of Pine, which is MIME-aware, and the next release, 3.92, can support 8BITMIME. If your mail server is running 8.7.3 sendmail, properly configured, you can also send 8-bit characters from programs like Elm without problems. There is another way to send 8-bit characters, which is limited to the headers of messages, and this is defined in RFC1522. It looks even uglier than the MIME QUOTED-PRINTABLE encoding that is common today, and an example of this can be seen in the Subject: header I have included with this message. There are presently very few programs which have this support, and Eudora seems to be one but with limits, Netscape has limited and buggy support, and there are a few others. Again, the next release of Pine3.92 will have this support as well, and I have been working with the authors to make sure the support is correct. If someone with a complete MIME-aware program were to view this message, it would appear as: v v , Subject: Cestina? Prosim!!! (compatible with non-MIME mail/news-readers, except Nyetscape or other programs which do not use fixed-width fonts :-) So a site needs to do the following for national support... o First, configure the system for the characters needed. Most sites already have this, although there are some which still suffer with stupid ASCII-only terminals. o Second, get a mailreader and newsreader which supports MIME, and preferably, also supports conversion from the local character set to the net-preferred ISO-8859-2. Sometimes, this could be done by dropping the DOS code page and using a code page which matches ISO-8859-2, as it is a question of finding a smart mail program or making the OS smarter. o Third, the mailreader and newsreader should have RFC1522 support to be able to read headers such as mine. o Fourth, to make things better for sites without MIME support and also to minimize the amount of QUOTED-PRINTABLE encoding that is needed, get a SMTP server with 8BITMIME support. As an example, a site running BSD Sendmail 8.7.3 on Unix with ISO-8859-2 fonts and Pine3.92 will have all the desired features. All these programs are available without charge (except for the hardware). Windows users and DOS users will have a bit more difficulty, but it should be possible to obtain ISO-8859-2-compliant programs. (One of my next steps will be to incorporate translation into programs, through GNU recode or similar.) I would like to see support for ISO-8859-2 MIME awareness and 8BITMIME mail as high as possible in SK/CZ and elsewhere in Europe. I had been thinking that I might volunteer to travel from institution to organization and volunteer to consult with the installation of needed programs to do this. Maybe I might do this in the spring and summer, if anyone is interested. ;-) This would also give me a better feel for what equipment is out there and what programs are used, so that I can make useful suggestions for developers of mail and news and other programs which would need to be localized for national support. I would rather see support everywhere in CZ/SK, and in the world, for mail and news in the native languages with complete accents, than to see US-ASCII continue to dominate, and I will be happy to help people who agree and want to use their native characters. However, I shall be conservative and write in ASCII in English here because I don't know Czech, although I always use German characters in mail sent to speakers of that language. vesel=E9 vianoce ;-) Barry Bouwsma, <barryb@tuke.sk> visiting TU Ko=B9ice, SK
Next Article (by Date): Re: Cestina??? Prosim, ne! Milan Zamazal
Previous Article (by Date): Re: Cestina??? Prosim, ne! Matus Uhlar
Articles sorted by: [Date] [Subject]