Since the distribution of document N1446 there has been only  a very limited amount of informed feedback on the original proposal and the US comments. There is however an increasing volume of software development that requires an accurate and comprehensive set of ISO standards. Much of this work seeks to restore the full richness of Armenian orthography, which is now well within the capability of modern information technologies. There is also a need to revise other ISO standards that will be derived from ISO10646. This discussion document  seeks to provide additional explanatory information, and to propose a workable reconciliation of the original proposal and the US ANSI feedback that permits the maximum restoration of Armenian orthography with minimal impact on existing applications.

In response to recommendations from the UNICODE consortium and the chairmen of ISO committees, the submitters of the original document have made recommendations to SARM (the State body of the Republic of Armenia responsible for Standardization ). SARM has upgraded its status with ISO so as to become a P member of TC46 and TC46/SC2, and is currently exploring the possibilities of becoming a member of ISO/IEC JTC 1 and ISO/IEC JTC1/SC2.

It may be useful to clarify that the original submitters, the Computer Standardization Committee of the Armenian Engineers and Scientists of America (AESA)  (hereafter referred to as AESA)  is an authorized agent of SARM and also act in an advisory capacity to SARM.  The AESA committee apologizes for any difficulties caused  by a misunderstanding of the term (requester type), and requests that the original document be understood as an expert contribution in accordance with the recommendation of US ANSI.  The AESA committee will be glad to use either a technical corrigendum or defect report in place of an addition form , if this is the preferred ISO mechanism for communication. This topic has been
discussed with Mike Ksar, convenor of the ISO/IEC JTC1/SC2/WG2 committee. AESA is prepared to follow his recommendations in this regard.  The AESA committee emphasizes its role as an advisory
body to SARM , which is the official state standards body of the Republic of Armenia and which is the final Armenian national source of authority on computer standards. This document reflects the views of the AESA committee that are not necessarily endorsed  or rejected by SARM, and merely seeks to clarify current open issues so that informed discussion can lead to informed decision making.

a) Implementation Level

 There is some semantic confusion that results from the UNICODE usage of "modifier letters" and the ISO10646 usage "implementation level". This is an area that requires the critical focus and attention of ISO and UNOCODE specialists if the full set of Armenian orthographic characters are to be correctly described in ISO10646. We draw particular attention to the pitifall of  misinterpreting correct usage and function from print samples that demonstrate little other than the electromechanical limitation of the printing device. We emphasize the importance of encoding both the full range of characters and encoding them in the correct manner in accordance with ISO  and UNICODE terminology, and are keen to disseminate full details of correct Armenian usage in order to achieve that objective.

The original request for level 3 implementation was based upon the description in section 15.3 of ISO10646 standard that prescribes that level 3 "MAY contain coded representations of any characters". This seemed advantageous , since "MAY" is non-definitive and flexibly, whereas level 1 is definitive and restrictive. We share the view of US ANSI that this is a substantial change, and understand the concern of
impacting current systems. However, we feel that this needs to be counter-balanced by the need to correct errors before they are perpetuated in future systems. We therefore urge systematic review and discussion of the differences of opinion so that a final resolution can be embraced that will serve future generations of
software developers and users.

b) Unification

The AESA committee would like to point out that significant effort has been expended on explaining and discussing the philosophy and ramifications of ISO10646 and UNICODE to scientific and governmental authorities in the Republic of Armenia, where they were considered with great interest and reviewed in considerable depth. In general, we are sympathetic to unification where it is appropriate, yet do not want to see specific features of Armenian orthography obliterated by a simplistic and naïve form of UNICODE philosophy, or an incorrect application of appropriate philosophy. If the symbols described are indeed missing in ISO10646, we are receptive to suggestions on the appropriate way to include them. The various suggestions that have been made to unify characters where there is no identity of both function and glyph form seem to be inappropriate applications , if not overt violations,  of  ISO10646/UNICODE philosophy.
I point out for the sake of those who lcak access to the UNICODE Standard 2.0 that that standard specifies
"Identifying a character A as a compatibility variant of another character B implies that generally A can be remapped to B without loss of information other than formatting."

In both formal and informal reviews and email dialogues, we have sought to explain the nature of the full Armenian orthographic system in order to facilitate correct  classification and unification decisions.. If  individuals familiar with ISO/UNICODE classifications  are prepared  to examine these symbols and the description of their usage closely, we are sure that the appropriate classification and status can be determined.

The following section is written to focus that necessary level of critical attention on the status of the specific characters.

Please note that the transliterated names are not being recommended for adoption for ISO or to codify any specific transliteration scheme.

We would like to achieve a definitive resolution (i.e. agree to agree or agree to disagree) on the place of these symbols in ISO10646/UNICODE so that applications development is not further restricted.

1)  Armenian MIJAKET

Description: MIJAKET is a separation mark used inside a sentence, and  literally translated as "middle dot". It is larger than the Armenian VERJAKET, has left alignment  and has a different function.

Original Proposal: We proposed that this should be added to at location 058A.

US ANSI comment: Unify with 00B& Middle Dot..

Discussion Comment: While this appears to be a useful and constructive suggestion, and there is some similarity of function between MIJAKET and the Middle Dot at 00B7, the symbol is nevertheless different and does not satisfy the criteria of common functionality and glyph form.  As a compromise, the US ANSI unification proposal could be accepted, albeit at the price of weakening ISO10646/UNICODE principles, since MIJAKET is not truly identical to Middle Dot as a character.

2 and 3)  ARMENIAN DZAKH and AJ CHAKERT

Description: (see original document).

Original proposal: A request was made for addition on the grounds  that (as implied in the names) that these are derived from other Armenian characters.

US ANSI Comment: Unify as glyph variants of of 00AB and OOBB. "We have seen  multiple instances of Armenian text using the regular quotation marks."

Discussion Comment: The last US ANSI comment seems to be an example of  the syndrome of accepting evidence from printing devices that lack the full range of symbols to argue that an incomplete set is complete. While there is a case for accepting unification solely on the basis of function, this does not contravert the argument  that
the character itself is not the same as the character with which US ANSI proposes unification. The unification proposal  is therefore a violation of ISO10646/UNICODE philosophy.

4) Armenian YENTAMNA

Description: "This is one of two hyphenation signs in Armenian…"  (see original  text  for complete description.)

Original Proposal: Addition due to lack of similarity to other characters.

US ANSI Comment:"Again we have seen evidences of usage of straight HYPHENATION (2010)  in Armenian text. The behavior of the Armenian matches the one expected by the regular HYPHENAATION character which is not linked to a specific script. Therefore we don’t see the need for the addition of this character."

Discussion Comment: The US ANSI comment is prone to the same logical flaw exposed above. This appears to be a violation of ISO10646/UNICODE philosophy, or at least a  poor application of it.

5 and 6) Armenian  MIUTIAN GTSIK  and Armenian ANJATMAN GITS

Descriptions: See original document.

Original Proposal: Addition

US ANSI Comment (abbreviated): "These are both explanatory marks and very similar in nature with the existing HORIZONTAL BAR (2015) and many other horizontal bars already encoded in ISO/IEC 10646. The standard does not mandate exact shapes of characters, nor does it specify character behaviour for symbols which can be used as explanatory marks. Therefore, before accepting these new characters, all effort should be made to merge them with symbols having similar shapes.".

Discussion Comment: This request could probably be accommodated as glyph variants  by unification of  Armenian Miutian Gtisk with 002D Hyphen, and Armenian ANJATMAN GITS with QUOTATION DASH (2105). They
still need to be added to the Armenian code page.

7 and 8)  Armenian SIUN and YERKVOREAK

Description: See original document.

Original proposal: Addition

US ANSI comment: Unify  SIUN with PRIME (2032) and YERKVOREAK with DOUBLE PRIME (2033)

Discussion Comment: Recommend  acceptance of US ANSI comment.
 

9)  YEV

US ANSI Comment: This is a partially documented addition without proper documentation, but since they are described in the original  proposal, we are commenting on them. (Other comments are made on nomenclature.) The US
is willing to consider inclusion of this character to provide a capital form of an existing small form.

Discussion Comment: This seems to be a positive resolution of a long-standing theoretical debate.  While YEV may be decomposed into two separate characters, it serves a function as the Armenian ampersand, and already exists as a separate typographic symbol. The US ANSI proposal is in accordance with the pragmatic approach adopted on recent private code page implementations.

10)  OU

Discussion Comments: This another problematic character from a variety of perspectives.  This character can be justified as an independent character on the basis of existing typographic practices.    It can also be decomposed and therefore considered as a digraph that does not need encoding.  Arguments about its inclusion or exclusion tend to be ideologically divisive due to variants in Soviet and Western Armenian orthography.
The US ANSI argument that it is similar to other digraphs is only partially valid, since those other digraphs (to the best of this author’s knowledge) may not have been used  (and taught in schools) as independent symbols. There is no easy decision on this character that can satisfy all parties. My personal preference is to include it on the grounds that any standard should serve as wide an audience as possible, and exclude as few developers and end users  as possible. Those seeking to exclude it should do so at the application level, and not from an international standard.

Five adjustments were proposed in the original document. We are quite content to resubmit these as technical corrigendum if that is the wish of the convenor  of the working group. These were originally grouped together since they are all combining characters that therefore require an adjustment to their
status property.

11 and 12)  Armenian DUR (0559) and PATIW(055F)

Everyone with any level of knowledge of Armenian characters recognizes that this is not used in any form of  Armenian orthography.  The US ANSI committee comment is that "the usage and meaning of the current character has always been unclear". We have yet to find any application that uses this character in any meaningful way.

Discussion Comments: Research indicates that this is most likely to be a misrepresentation of a pair of Armenian symbols (DUR and PATIW) that both need an glyph adjustment to reflect their correct form. The US ANSI comment that this be corrected  by  a technical corrigendum  is very acceptable.

13, 14, and 15) Armenian SHESHT, BATSAKANCHAKAN  and  HARTSAKAN

Original Proposal: Change the property  of these characters so that they become combining characters and are handled correctly by font manufacturers..

US ANSI Comment: Make any necessary changes using a technical corrigendum.

Discussion Comments: No further discussion required,. Implement in the manner acceptable to committee convenor.

a)  Name Changes

We have tried to avoid unnecessary controversy and discussion of transliteration issues in a group dedicated to character set issues. We accept  the US ANSI recommendation that naming issues be handled
in Annex P and any other pertinent forum (e.g. TC46).

b)  General Policy

We support the policy of avoiding unecessary changes. We understand and share the concern of the US ANSIT committee that risks to existing implementations be minimized by making additions rather than changes. However that risk needs to be evaluated against the restrictions on current and future growth
that result from errors and omissions. We support the serious efforts of those members of the international
software development and standards committees who seek to establish, robust, reliable and correct standards, and hope t hat this document will lead to the endorsement of the optimal set of solutions.