From keld@dkuug.dk Wed Feb 27 18:21:50 1991 Received: by dkuug.dk (5.64+/8+bit/IDA-1.2.8) id AA23050; Wed, 27 Feb 91 18:21:50 +0100 Date: Wed, 27 Feb 91 18:21:50 +0100 From: Keld J|rn Simonsen Message-Id: <9102271721.AA23050@dkuug.dk> To: erik@sra.co.jp, wg15rin@dkuug.dk Subject: Re: (wg15rin 80) national profiles X-Charset: ASCII X-Char-Esc: 29 > Is Japan expected to produce a sample Japanese national profile? Or > should Japan produce The Japanese National Profile? You should try to do both! But for WG15 purpose, you should only try to do a sample national Japanese profile. This is at least what was told to Danish Standards for the profile to be included in the POSIX.2 standard: it must only be an example, bearing no formal weight. This is also easier for you, as you are allowed to make errors... In the end I would expect you to have a need for more than one national profile, that is at least what we have in mind in Denmark... > According to the London minutes, Japan seems to have been given an > action item to produce a document about locale harmonization. Is this > the same as guidelines for national profile writers? As far as I remember, yes. > The Danish sample national profile appearing in ISO POSIX.1 contains > references to the sample charmap and locale in POSIX.2. However, if a > national profile is something that must be conformed to, is it > reasonable to specify a charmap? The boundaries between POSIX.1 and POSIX.2 are not fully clear. We were asked to provide an Input Method in POSIX.1 and you know Input Methods! We did an IM dependent on the ISO_10646 charmap. That should be general enough to us. > My current thinking is that in order to promote interoperability and > application portability, we will want to produce national profiles > that serve as the basis for conformance testing. We can put a charmap > or two in informative annexes separate from the profiles themselves. > > For example, in Japan's case we might specify the JIS X 0208 > repertoire in the national profile itself, and provide a sample > Shift-JIS charmap in an informative annex. Well, I think the 0208 and the Shift-JIS charmaps should contain the same symbolic character names when the characters are the same. 10646 should contain all of them. Thus it should be safe just to specify one general charmap, the 10646, and all other should be consistent with this. For conformance testing, well you need to be able to test everything. As we have sepcified things, you are able to specify the full 10646 repertoire in any (decent) character set, containing the repertoire of invariant 646. Keld