Pd: Odp: Re: Unicode fundamental character identity

[piotrunio-2004@wp.pl]

Dnia 01 lutego 2025 00:48 James Kass via Unicode <unicode at corp.unicode.org> napisał(a):  On 2025-01-31 11:15 PM, piotrunio-2004 at wp.pl via Unicode wrote:   Dnia 31 stycznia 2025 23:45 James Kass <jameskass at code2001.com>  napisał(a):      On 2025-01-31 9:28 PM, piotrunio-2004 at wp.pl via Unicode wrote:          The proposal L2/25-037 already shows a difference in plain         text of the         HP 264x characters, where 0x12 (2) connects below vertical or         perpendicular diagonal, whereas 0x18 (8) connects below         diagonal of         same direction. Those are different types of connections which         is a         plain text distinction of box drawings.      A "smart" font dedicated to these characters would provide appropriate     glyphs based on context.  This would result in a plain-text display     identical to the original display.   That doesn't make sense because on a fundamental level, in a legacy  computing semigraphical environment, each character tile is drawn  independently, and only affects the area of the screen dedicated to  that character. Having a context dependent system would overcomplicate  the renderer beyond the scope of the original system. Furthermore, on  the HP 264x system, the two characters can exist in isolation (as  shown in obGQ4Ie.png (1440×720) (imgur.com)  < i.imgur.com https://i.imgur.com/obGQ4Ie.png>),  and the user can in fact type the  two characters differently, with the 2 and 8 keys as shown in page 31  of 204 in  02645-90005_2641A_2645A_2645S_N_Display_Station_Reference_Manual_Nov1978.pdf  (bitsavers.org)  < www.bitsavers.org http://www.bitsavers.org/pdf/hp/terminal/264x/2645A/02645-90005_2641A_2645A_2645S_N_Display_Station_Reference_Manual_Nov1978.pdf>.   Sorry for the confusion.  I'm referring to a Unicode "smart" font  working on a modern system displaying Unicode plain-text.  This is all  automatic and handled by the rendering system.  If a dedicated font is  used to display the text, contextual glyph substitution would make the  display indistinguishable from the original display on the legacy  system.  Also, on a modern system any "dumb" font supporting the  characters would still produce a *legible* display, even though it might  not be as pretty.  And legibility in plain-text is one of the factors  driving encoding decisions.  (This might be why font selection was  mentioned as a solution in the document referenced earlier.)   That still cannot possibly work on isolated instances of the characters. In fact, if you have two different Large Character set strings that only differ by the use of 0x12 or 0x18 character, then the HP 264x will display them distinct in Unicode 16.0 mapping they will result in the exact same string and no amount of contextual glyph substitution will work. And as I said, complex features such as contextual glyph substitution are fundamentally completely out of scope for characters that originated from semigraphical text, no matter how modern the system displaying it is.           Data loss in round-tripping is implicitly evident from the         information         provided in the proposal: if an HP 264x Large Character set mode         document has the characters 0x12 0x18, it converts to Unicode as         U+1CE2B U+1CE2B, which converted back to HP 264x Large         Character set         mode is 0x12 0x12, which loses the distinction between the two         characters and will appear slightly differently than the original         document on HP 264x platform.      Yes, this is implicit in the proposal.  Any future proposal should     make     it explicit while referring to the earlier proposal for background.     Please keep in mind that the committee members must wade through many     different proposals covering all aspects of character encoding.      Keep it     short, straightforward, and simple as possible to ease their burden.   The character has already been proposed. What would any future  proposal have to do with that?   If my understanding is correct, the character has already been proposed  and rejected.  It's not uncommon for a subsequent proposal to be  submitted which addresses concerns raised during the rejection of an  earlier proposal.  (If my understanding is not correct, someone will  probably set me straight.)   It's not in  www.unicode.org Archive of Notices of Non-Approval (unicode.org)  so it's not rejected.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://corp.unicode.org/pipermail/unicode/attachments/20250201/a314968c/attachment.htm>