RE: Hex-byte pictures (WAS: RE: Hexadecimal digits?)
Hi Philippe! When dealing with protocol specifications, there's often a need for characters like these, too, since hex byte pictures are unambiguous. I have a DEC dumb terminal around here somewhere which also uses them when debugging control characters. I suppose you could argue it's purely a formatting issue, though. If you've got some technical documentation reference of this terminal, it would be worth to give it as it will be used in technical documentations. It's a DEC VT320, and it's second hand like all of my dumb terminals, so I've never actually had the original manual. Upon closer inspection, it only appears to do hex-byte pictures for some C1 control pictures -- see http://vt100.net/built-in_glyphs.html. The VT220 did a similar thing, but more of it - no cuddly names for NEL and so on, plus some other chars have hex-byte pictures (probably as they were unassigned, but I am unsure) -- see http://vt100.net/docs/vt220-rm/table2-16b.html. I'm pretty sure my Wyse WY60's (and probably my WY85's too) do the same thing, but they're so buried under junk it's probably not worth pulling them out to check. What you suggest is something else: it's a proposal to encode technical characters similar to control images, or to glyphs of keys on a keyboard. It is not a script, but a handy collection of unique glyphs. I feel we're on the same wavelength now! :) Indeed, not a numeric system but technical symbols. In a similar technical domain, I don't know if the technical glyphs that are (were?) used on terminals for IBM MVS systems, are all encoded. I remember there was a sort of zig-zag arrow pointing to bottom left, as well as other symbols denoting the current state of the terminal, and a few others to denote editing operations in a screen mode: one had to mark a edited line with a symbol, and the terminal took care of remember where editing was allowed and performed, and once you had created a modified line, you pressed a Send key to get the screen updated with the new text after editing operations. Sounds very familiar :) Stuff like the stick figure (which on some terminals looked more like a cowboy), don't appear to be in unicode, but then again, were those characters ever actually a part of the IBM 3270 charsets, or were they simply internal only? This was more or less working in a way similar to the vi editor line-mode interface, except that it was screen-based rather than line-based. Looking at the original proposal by Frank da Cruz again after so long (http://www.funet.fi/pub/kermit/ucsterminal/hex.txt) reveals it cites many documents. Have a look at http://www.funet.fi/pub/kermit/ucsterminal/terminal-exhibits.pdf (~2.7MiB). BTW, Frank also had other proposals which included the IBM 3270 characters I think you were referring to (poke around the directory at http://www.funet.fi/pub/kermit/ucsterminal/).. I like the hex byte pictures proposal, and I'm seeing more reason to like it, the more I look into it.. Cheers! - Simon
RE: Hex-byte pictures (WAS: RE: Hexadecimal digits?)
BTW, Frank also had other proposals which included the IBM 3270 characters I think you were referring to (poke around the directory at http://www.funet.fi/pub/kermit/ucsterminal/).. I am not proposing to encode all terminal function indicators in Unicode. Else it would mean that we could as well standardize of those many icons found in toolbars of GUI applications. If these symbols appear in printed documentations, they are still snapshots of icons used in a particular program or device, and not intended to be part of interchanged documents... snip I wasn't either - just pointing it out. I agree entirely :) - Simon
RE: Hexadecimal digits?
Hi :) http://www.dkuug.dk/jtc1/sc2/wg2/docs/n2677 N2677 Proposal for six Hexadecimal digits Ricardo Cancho Niemietz - individual contribution 2003-10-21 snip Could be interesting for processing, and I can see a reason for keeping these unique from U+0041-U+0046 but ultimately I thought the hex byte picture proposal would have been more useful. - Simon
RE: Hexadecimal digits?
Hi Philippe, http://www.dkuug.dk/jtc1/sc2/wg2/docs/n2677 N2677 Proposal for six Hexadecimal digits Ricardo Cancho Niemietz - individual contribution 2003-10-21 snip Could be interesting for processing, and I can see a reason for keeping these unique from U+0041-U+0046 but ultimately I thought the hex byte picture proposal would have been more useful. Why that? How will you represent hex sequences with variable number of nibbles? The purpose of this proposal is to make those extra characters really numeric and not letters, with only a compatibility equivalence (not a canonical one) with ASCII letters to which they ressemble. snip Of course, and I agree with you entirely when you're dealing with an arbitrary number of nibbles, and I support this proposal for the same reasons you do. However personally, when dealing with a octet, or an arbitrary number of octets, I believe the byte-pictures would be much easier to deal with (especially when dealing with a lot of raw data). - Simon
RE: Hexadecimal digits?
Hi Philippe, However personally, when dealing with a octet, or an arbitrary number of octets, I believe the byte-pictures would be much easier to deal with (especially when dealing with a lot of raw data). Except that it would require 256 new codepoints, instead of just 6 for the proposed HEX DIGIT characters. What is complicate, when dealing with lot of raw data, to convert it to nibbles then coded with numeric code points, rather than converting bytes to code points? You just add a shift and mask operation to output 2 code points rather than just adding each byte as an offset of a base code point. Still, you need to convert your raw data to suitable code points to display the HEX BYTE characters. snip I never said there was anything complicated about it, I said I personally prefer the hex byte characters - They're a much more compact and elegant solution to representing octets. When dealing with protocol specifications, there's often a need for characters like these, too, since hex byte pictures are unambiguous. I have a DEC dumb terminal around here somewhere which also uses them when debugging control characters. I suppose you could argue it's purely a formatting issue, though. What you propose is NOT a complementary set of digits for base 16, but a complete new set of numbers in base 256, so that a glyph like [00] will be displayed instead of just 0 (this is a disunification of all the existing ASCII digits, as if it was a new script using its own numbering system)... snip Well I didn't propose it, but I do like it! :) Other historic numbering systems are used today and better suited for representation, notably the compound base (12, 5), when people where counting the first digit in one hand with the first finger pointing on the 3 phallanges of the 4 other fingers, and the other hand was used to count the second order digit by raising each of its 5 fingers. snip I do not see how historic numbering systems are appropriate for representing octets, which was the point of the proposal. I strongly doubt the Babylonians or the Mayans considered computer engineers would settle on 8-bits to a byte with base-60 or base-20 respectively. I'm not sure what you meant by most of your message, though. I'm talking about representation, in a similar vein as the control pictures section (U+2400-243F), and not a numeric system. - Simon
RE: Traditional dollar sign
Hi! snip However, the presence of two opposing conventions serves as a strong hint that there was no consensus in 1966, nor now, as to how glyph variants of the dollar sign were to be used to stand for different types of dollars. I went to school in the 1980's, and both in Victoria and Tasmania I was taught to write it using the double-bar form. My brother in law is a school teacher here in Victoria and says he's been told to teach kids to write it using the double-bar form in Victoria and New South Wales, and strongly discourage the single-bar form. He doesn't know about other states. Kevin later quoted the Decimal Currency Board: (c) where it is necessary to distinguish the Australian dollar from overseas currencies, the letter A should be placed immediately after the dollar sign - $A; Interesting. I've often seen the opposite, A$ or AU$, even in contexts that only involved Australian dollars, not U.S. dollars. Of course you can always just use AUD and USD and be done with it. My bank (ANZ) recently gave me literature related to obtaining foreign currency, and used the form $A (that is, with the double-bar form of the dollar sign, not the single-bar form). Considering the small glossy leaflet was about the rising Australian dollar, it's evidently a recent publication. Their website, however, obviously has no choice but to use the single-bar form due to font authors, who appear to be quite consistently using the single-bar form. Curiously, though, my bank statements from ANZ use this single-barred dollar sign ;) Considering recent publications, the site pasted (thanks to Kevin Brown), the wide knowledge of the (original) double-bar form of the dollar sign, and the fact that it's still taught at schools in Victoria, Tasmania, and New South Wales (possibly other states too, I'm unsure) - does this amount to reasonable evidence of an existing subset of users? The use of the single-bar dollar sign on the website Kevin provided (http://www.abs.gov.au/Ausstats/[EMAIL PROTECTED]/0/c7103f5100c7663fca2569de00293f3c) is obviously because there's no reliable method of displaying the double-bar form! I smell a subset of users which would benefit from disunification right there ;) - Simon
Traditional dollar sign
Hi! Just a quick question.. The description for U+0024 (DOLLAR SIGN) states that the glyph may contain one or two vertical bars. Is there a codepoint specifically for the traditional double-bar form, or any plan to include one in the future? I was taught at school that the double-bar form was used when Australia switched to decimal currency in 1966, and that it was incorrect to write the single-bar form when referring to Australian dollars. I guess the single-bar form had taken over due to the lack of support from type-faces and computing devices, although it's still quite common to see it in Australian publications, especially in large fonts (headlines, advertising, etc). Cheers! - Simon
RE: Traditional dollar sign
Hi! snip I was taught at school that the double-bar form was used when Australia switched to decimal currency in 1966, and that it was incorrect to write the single-bar form when referring to Australian dollars. It would be interesting if you could document that. That could be tough :) Literature shown to me was at school (many years ago), and digging it up would be difficult. It's widely known that the double-bar form does exist, though, at least! I guess the single-bar form had taken over due to the lack of support from type-faces and computing devices, although it's still quite common to see it in Australian publications, especially in large fonts (headlines, advertising, etc). It looks like actual practice is what you describe: the free alternation between the form without change in meaning. If we were to add a code point we would get into the situation that the free alternation would suddenly become a matter of content difference (not just a choice in presentation). In other cases where the majority of users freely alternate, but there is indication that some subset of users need to maintain a form distinction we have used standardized variants. This has been done mostly for mathematical symbols. snip I understand, although couldn't that same argument be used against many of the characters in the 'Dingbats' section, such as the ornamental variations of exclamation marks, quotation marks, and so forth? I do realise these come from an existing character set, but there are indeed still users of the double-bar form. Even my Concise Oxford Dictionary is printed using the double-bar form (under the term, 'dollar'). I just thought it extremely odd that a character which is still in common (albeit admittedly waning) use is not included in the set. Peter Kirk made a valid observation with regards to the Lira symbol (U+20A4) which Unicode admits often has U+00A3 (Pound sign) used in its place, with the only difference being a double-bar on U+20A4. Cheers, - Simon