We have a problem...
In PKIX part 1, chapter 4.1.2.4 "Issuer Name" the issuer field
(and the subject field similarly) is defined. There are no
mandatory X.520 attributes (like country or commonName etc)
but the the DirectoryString is defined in ASN.1 as:
> DirectoryString ::= CHOICE {
> teletexString TeletexString (SIZE (1..maxSize)),
> printableString PrintableString (SIZE (1..maxSize)),
> universalString UniversalString (SIZE (1..maxSize))
Below this definition it is stated that:
> The directoryString is defined as a choice of PrintableString,
> TeletexString and UniversalString. Conforming CAs shall choose from
> these options as follows:
>
> (a) if the character set is sufficient, the string will be
> represented as a PrintableString;
>
> (b) failing (a), if the teletexString charater set is sufficient,
> the string string will be represented as a TeletexString;
>
> (c) failing (a) and (b), the string shall be represented as a
> UniversalString.
The problem for us outside the US is how to encode national characters.
In Sweden for example, PrintableString which is equivalent to 7-bit ASCII
is not sufficient. According to the standard we would then make use of
teletexString which is equivalent to the T.61 character set.
Now, T.61 is an old standard that is very seldomly used nowadays. In
particular neither Netscape Navigator nor Microsoft Internet Explorer
make use of T.61.
Instead both Navigator and Explorer uses the character set in ISO 8859-1
(Latin-1 which is equivalent to 8-bit ASCII). This character set is used for
both the PrintableString tag and the teletexString tag.
Clearly the use of ISO 8859-1 in conjuction with the teletexString tag is an
error and some X.500 directories won't accept this. The effect would be
that searching these directories becomes impossible.
The other alternative (ISO 8859-1 with the PrintableString tag) looks very
attractive IMHO and I would prefer that this solution were incorporated into
the standard. There are still problems however: many complilers based
on ASN.1 doesn't accept 8-bit characters in a PrintableString.
Can anyone on this list see a solution to this problem? If so, I would suggest
that the PKIX-1 draft is updated accordinly.
Regards,
/Lars Johansson
Sweden Post