1 17 June, 1996 Title: Proposed Resolution of Japanese Comments on JTC 1 N3745 Source: Edwin Hart and Alan Griffee (Acting Editors) Action: For the consideration of the Japanese National Body and WG 2 Distribution: Japan, ISO/IEC JTC 1/SC 2/WG 2 We thank the Japanese National Body for the comments on JTC 1 N3745 on the NP for a Technical Report on the topic, "An operational model for characters and glyphs". This document responds to those comments. Note that the acting editors appreciate the additional information about the Japanese concerns obtained through informal discussions with Mr. Takayuki K. Sato. Comment on Question 1 The title and scope of the project should be limited within Latin based Scripts(s). At least it should not include CJK Ideograph and Graphic symbols. Proposed Change: Title to be: An Operational Model for Characters and Glyphs for Latin-Oriented Scripts. Scope to be change in line with the changed title. Rationale 1. The draft TR does not address the "variety of shapes issue" for CJK ideographs. [rephrased] Japan is concerned that the draft TR presents an idealized view of characters and glyphs that conflicts with the rules used to unify the ideographic characters in ISO/IEC 10646-1:1993. Ideally, a character embodies logical information and a glyph embodies just the shape. SC 2 encodes the logical information of characters into its standards. However, the unification rules used the shape of ideographs, rather than the information content, to unify the ideographs for coding into 10646. Should unification of ideographic characters be based on the information content or the shape of the character? Since these issues are ignored by the draft TR, Japan requests that the TR be retitled and that the scope be changed to reflect this concern. Rationale 2. The symbols encoded in ISO/IEC 10646-1:1993 are the result of unification by their shapes [rather than their information content]. Mr. Sato cited an example where Japan requested that the pictorial representations of the control character symbols from ISO 2047 be added to 10646. Japan thought that these symbols were presentation (glyph) variants and therefore should be unified with the alphanumeric representations encoded at U+2400 to U+2421 or encoded contiguously in this order. Instead of organizing and unifying with the alphanumeric representations of the control characters at code positions U+2400 to U+2421, WG 2 decided to unify them with symbols scattered in the symbols area of 10646. This is an example where the characters were unified and encoded by shape (glyph) rather than by the information content. Japan is concerned that this conflicts with the principles proposed in the draft TR. Response Japan is correct that the paper does not attempt to address the "variety of shapes issue" for CJK ideographs or unification of symbols by shape rather than information content. Although the intent of the draft TR was to address the character and glyph issues for all of ISO/IEC 10646, we have failed to include CJK ideographic issues. In addition, we should note that the draft TR represents an idealized view of characters and glyphs. ISO/IEC 10646-1:1993 includes characters from standards that predated 10646 and this draft TR. Thus, the developers of 10646 and the developers of the pre-existing standards did not have the benefit of this idealized view to decide characters from glyphs. Moreover, even if the developers of 10646 had had the benefit of the draft TR, compatibility with pre-existing standards far outweighed these idealized principles. We agree that the TR should not be delayed to include these topics and to change the scope. However, we recommend keeping the original title. Although the draft TR emphasizes examples from the Latin script, it also includes examples from the Arabic and Devanagari scripts. (See page 7 of the 11 June 1995, working draft.) Therefore, given the original intent of the document to address all scripts and that the draft TR includes examples from other scripts besides Latin, we would like to suggest a different solution and hope that it satisfies the Japanese concerns. We suggest (a) adding a disclaimer in the scope of the document but (b) keeping the original title. We also suggest that in view of the principles in the draft TR, Japan may wish to resubmit a new request to unify the pictorial representations of the control characters of ISO 2047 with the alphanumeric representations in ISO/IEC 10646-1:1993. Here is the suggested text to resolve the Japanese concerns. Replace the first paragraph of the "Scope" section: The purpose of this Technical Report is to provide an idealized framework for discussing characters and glyphs. For illustration, the Technical Report uses examples from scripts encoded in ISO/IEC 10646 and glyphs registered according to ISO/IEC 10036. The Technical Report addresses the following topics: Add the following to the end after the bullet list: The Technical Report avoids describing the complex issues associated with unification of the ideographic characters used in Southeast Asia. Note: This Technical Report describes idealized principles that may not necessarily be achievable in practice and were not completely followed in encoding characters for ISO/IEC 10646 and in registering glyphs according to ISO/IEC 10036. The fact that ISO/IEC 10646, ISO/IEC 10036, and other standards do not completely follow the principles in the model does not invalidate the model and does not diminish the utility of having the model. We hope that the proposed resolution and additional text satisfy the Japanese concerns. If they do not, please tell us your concerns and any particular suggestions for addressing those concerns in time for us to respond by the SC 2/WG 2 meeting in August. Electronic mail and facsimile information are given below: Edwin Hart: edwin.hart@jhuapl.edu +1.301.953-1093 (facsimile) Alan Griffee: afii@ix.netcom.com +1.303.924-9912 (facsimile)