L2/03-090 Title: Compatibility Decompositions for Squared abbreviations Author: Asmus Freytag Action: Input to UTC Unicode 4.0 adds a few more squared Latin abbreviations that brought with them new compatibility decompositions. This document is intended as a resource for UTC in reviewing the proposed decompositions. Usually, we spell out the glyph for the 'squared' Latin Abbreviations when we do our decomposition. Sometimes the capitalization of the compatibility character is different from the common appearance of the abbreviation in normal text. This can lead to issues with decompositions. The following is a brief rundown of past and new characters that had glyph or decomposition issues. Two letterlike characters TEL and now FAX, are considered squared abbreviations, even though they are not named that way and don't occur in the CJK Compatibility block. 2121 TEL glyph and decomposition changed from T E L to T E L in Unicode 3.0 2222 FAX proposed decomposition is F A X In both cases, presumably, we intend that other capitalizations can be unified with these (eg. Tel or Tel. etc.) 32CF LTD (glyph is in caps) prosed decomp L T D The all-caps use is not the most common. If we ever get a Ltd or Ltd. compat characters (that we feel we need to support) we may need to decide whether we should clone the character. That's what we did for 339E km and 33CE KM however, in the latter case the SI has specific rules for capitalization which are lacking for the commercial abbreviation. A similiar issue holde for 3XXX PTE (glyph is in caps) Both LTD and PTE were added to accommodate the DPRK. 3396 ml This has 2113 as the 'l' in glyph and decomposition. However, the glyph changed from Ml to ml between 1.0 and 2.0. Curiously, the glyph for this character in MS Mincho is still Ml, which is the reason that it is wrong in the 4.0 beta charts. (in the SI system 'M' is Mega, and 'm' is milli. Even though 2113 is not (or no longer?) an SI symbol for liter, Ml with capital M would presumably mean Megaliter, an unusual quantity. It's speculation whether the glyphs for MS Mincho were designed based on Unicode 1.0, or whether there really is a strong tradition for using a capital M. 33D7 pH has a historic decomposition as PH, which is unchanged, even though the glyph was changed from PH in 1.0 to pH in 2.0 and has remained that way and is now pH in Mincho. Since we did change the decomposition for TEL, the retention of this unusual decomposition in 2.0 should be considered a defect.