ietf-822
[Top] [All Lists]

Re: 8-bit transmission in NNTP

1994-09-25 20:09:36
Masataka Ohta writes:

If ISO-8859-1 is displayed on equipment without 2022 support, the message
becomes non-intelligble.

Nah, that depends. In many cases the equipment will have ISO-8859-1
support, and then the message is OK. Anyway 2022 support is not
significant here. 

Urrrrr, perhaps, you are assuming that ISO 8859/1 is used only in 8bit
environment and left/right parts of ISO 8859/1 are designated to GL/GR.

But, as I pointed out earlier about ambiguity of ISO-8859-1, that's
merely a unfounded assumption (be sure that ISO-8859-1 and ISO 8859/1
are different things).

And ISO-8859-1's assumption on GL/GR designation is based on ISO 2022.

Thus, if ISO 2022 is not supported by a terminal at all, ISO-8859-1 nor
ISO-2022-INT-* can not be displayed.

Of course, to display ISO-8859-1 or ISO-2022-INT-*, full support
of ISO 2022 is not necessary.

If mnemonic is displayed on equipment without 646 support, the message
becomes non-intelligble.

In the internet you can assume 646,

As far as MIME concerns, we are not assuming 646.

From RFC 1521;

      NOTE: This subset has the important property that it is
      represented identically in all versions of ISO 646, including US
      ASCII, and all characters in the subset are also represented
      identically in all versions of EBCDIC.

as the internet mail rfcs
require support for 646 (I assume you mean ISO 646, which is
almost the same as ASCII, or US-ASCII as it is known in MIME.)

As RFC 822 requires support for US-ASCII, including NULL byte, I
assumed US-ASCII (but not NULL) to designed ISO-2022-INT-*.

So that is the difference: on the Internet you can always assume
to have the nessecary hardware support for mnemonics, which is not the
case for the two other examples you mentioned.

I can't understand what you want to say here. What is "hardware support"?
What is "two other examples"?

That is,
when the 2022 shifts are not honored, the message is displayed
most likeky in another character set, and the characters shown will
most likely have no correlation to the ones intended. 

The situation is no different for ISO-8859-1 or mnemonic.

The situation is different for mnemonics, which only assumes ASCII
support.

OK. ISO-2022-INT-* also assumes ASCII support only for the transport.

And, on pure ASCII terminals, both your mnemonic and ISO-2022-INT-*
displays pure garbages on most non-English languages.

Mnemonic might be a useful convention in a small community with European
culture in a special environment where 646 IRV are widely supported but
2022 are not. And even then, viewing it is quite uncomfortable.

I agree that viewing mnemonic is uncomfortable, as what you really
want to see is the real characters.

No. your mnemonic is uncomfortable for a small community with European
culture. Outside it, it is just unacceptable.

Don't forget that, your mnemonic for Han characters is just another
numeric encoding. Even for non-Han Japanese characters, your mnemonic
is unreadable unless I had some knowledge on the encoding. To make the
matter worse, the encoding is quite lengthy and inefficient.

I'll be a lot more comfortable to see 'A' as 65 (decimal ASCII value)
than seeing mnemonics.

A number of schemes can be made
to have this working, without problems. The problem only occurs when
the right hardware and software combination is not available.

Then, what's wrong with QP?

Any encoding which can not co-exists with ASCII (because of '=' for QP)
is unusable on local operating systems without file types.

This is unfortunately the case on the internet in many circumstances.
The problem is then what to do about it, and how to avoid the
problems when specifying a new standard for netnews.

I am arguing that using mnemonic encoding on the wire will remove
or substantially lessen some problems we have seen with MIME technology
in Internet mail.

My name is

        ^[$(_at_)B@ED!!>;9'^[(B

with ISO-2022-INT-* encoding.

Could you show how you can use your mnemonic on my name to remove or
substantially lessen some problems we have seen with MIME technology
in Internet mail?

If you can't, I'd like to conclude:

   Mnemonic might be a useful convention in a small community with European
   culture in a special environment where 646 IRV are widely supported but
   2022 are not. And even then, viewing it is quite uncomfortable.

Outside the community, the resulting encoding is completely unacceptable,
because we already have a lot friendly ASCII-transliteration systems of
our own languages.

Yes I understand that there are such systems available in japan,
and other east-asian countries using Han characters.

I also understand that some use of 2022 encoding is widely implemented
and I see that as a good thing. From the info I have got on this,
I would assume a 2022 utilizing encoding would be the most preferrable
in these cultures. It works well, and there is no sense in breaking
this practice. Adding MIME-labelling would be advantageous, tho.

I'm not objecting to add MIME-labelling of ISO-2022-INT-[123...], if it
is advantageous, that is, during the transition phase from mere
localization to full internationalization.

Also using mnemonic for characters not representable in these encodings
would be a plus, which would not harm anybody out there.

Ambiguity with '&' is quite harmful.

As far
as I understand it is not possible to write my name or postal address
in many of the far-east encodings. Using mnemonics you could.

Your name, "Keld" can be represened in all the two byte encodings.

JIS X 0212 now includes a lot of diacriticized Latin alphabets.

BYt, why you insist on "far-east encodings", when I'm proposing fully
internationalized encoding?

ISO-2022-INT-* can include all the ISO 8859 families.

Anyway European culture/language is quite widespread in the world,
it is dominant in Europe, the Americas, Australia, and very frequent
on most other continents. As such you cannot say that it is a small
community, that a quite readable fall-back scheme like mnemonics
is good for.

I only claimed that the society for which your mnenomic is useful, is
small.

   Mnemonic might be a useful convention in a small community with European
   culture in a special environment where 646 IRV are widely supported but
   2022 are not. And even then, viewing it is quite uncomfortable.

                                                        Masataka Ohta

<Prev in Thread] Current Thread [Next in Thread>