From alb@sct.gouv.qc.ca Tue Feb 4 20:20:36 1997 Received: from socrate.riq.qc.ca (socrate.riq.qc.ca [199.84.128.1]) by dkuug.dk (8.6.12/8.6.12) with SMTP id UAA21916 for ; Tue, 4 Feb 1997 20:20:29 +0100 Received: from 506.riq.qc.ca (riq-44-183.riq.qc.ca) by socrate.riq.qc.ca (5.x/SMI-SVR4) id AA22458; Tue, 4 Feb 1997 14:23:42 -0500 Date: Tue, 4 Feb 1997 14:23:42 -0500 Message-Id: <9702041923.AA22458@socrate.riq.qc.ca> X-Sender: alb@riq.qc.ca X-Mailer: Eudora Light pour Windows Version 1.5.2 Mime-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 8bit To: trondt@barsek.hsf.no (Trond Trosterud) From: "Alain LaBont/e'/" Subject: Re: URGENT on keyboards Cc: everson@indigo.ie (Michael Everson) At 16:10 97-02-04 +0100, Trond Trosterud wrote: >Dear Alain! > >Our keyboard lauout discussion carries on. > >After we talked I got a comment from a programmer in Finland, with >experience in doing Sámi keyboards for Dos and different Windows >applications. His points are the following: > >1. ALT is a 'syskey' in Win3.1x and Win95 and should never be used >2. AltGr is in use in Word, AltGr-T gives the TM-sign >3. Ctrl is in use in Word and Pagemaker as a special key >4. The deadkeys D12 and E12 (using 9995 definitions) should not be used, as >they are needed as dead keys. If Windows is installed with an american >keyboard it is OK to use them, but we cannot be sure that the installation >is carried out in this way. > >1. and 3. we can disregard. Noone will touch ALT or Ctrl for character >purposes (at least not to my knowledge). > >To 2. I have got the comment that TM will have to be taken from the Symbol >menu, and to 4. that the problem can be solved. About (2), WinWord is full of those flaws which can be short-circuited and vary from national version to national version (the flaws), as if they would not have known that they themselves supported keyboards that are incompatible with their goodies (for example, I can't enter French quotes using my keyboard alongside with my Canadian French version of Word! and as I write in en and fr, this causes me enormous problems, even if I short-circuit their smart quote function - the French left quote is stil unavailable, I have to cut and paste!) I keep complaining and they should see a copy of this (they are bcc'ed) I don't understand your problem with (4). I use this key both as a dead key and as a non-dead key on my national keyobard (in group 1 it is not a dead key, in group 2 it is one). "Dead keys" are not hard-wired, they are, like any other key, fully programmable. > >What is your comment? I have an awkward feeling I have cheated this >discussion into a dead end (we should have stuck to the minimum glyph >number possible). Can you save me out of this? How does the Canacian >standard look like? How do you manage? Have a look at http://www.olf.gouv.qc.ca/techno/images/clavier4.gif It is hard to see, though, but it will give you a firts-sight idea. We manage quite well with it, except a few flaws in WinWord taht I hope will be fixed over a short period of time. >What Level 3 positions are the most >secure ones (given that no Windows programmer respects a 3rd/4th level for >glyphs only as of today)? That is not correct. Only a few programmers nowadays make flawed programs that make assumptions on the national keybaord and these programs will inevitably have to change in an international market, no 2 keyboards in different countries are the same and level 3 is certainly in use in all European countries (300 million people at least). That said, there is a slight management problem we have in Canada with IBM keyboards. Our standard supports both 47- and 48-key alphanumeric keybaords and the scan code associated to C12 on a 48-key keyboard (Eurpean-style) is the same as the one used on D12 on an American-style keyboard. In the Québec government we have circumvented this in standardizing on 48-key keyboards and mandating them. We don't want to hear about implementation of the Canadian standard on a 47-key American-style keyboard which lacks one key (the ùÙ¦ characters; on this 47-key version the ùÙ can be obtained with a dead grave key in group 1 just for this reason, otherwise that dead key is useless for us, all àÀèÈùÙ are available in fully-formed characters on the keyboard; the ¦ is not available at all on a 47-key version, but this character is pretty unused so that is not a big issue although there is an issue!) > >These questions are the important ones. The rest is FYI: > >I still would like to see the ISO 14something standard on keyboards, if >possible (it might be on its way). I send it to you... you should have been able to get a copy before publication from the Finnish national body, but I can provide you one now. > >In my opinion, we should have a keyboard standard structure like this: > >K1. >- An OVERALL option for typing every single glyph of BMP decimally >- An OVERALL option for typing every single glyph of BMP hexadecimally >(I do not quite see which one of the two should be dispensed with, a simple >macro would give us both) I agree. Unfortunately the UCS only documents hexadecimal values, even if HTML 3.2 decided to use decimal values only. > >K2. >- An easy-to-remember key oriented way of typing all latin-based glyphs of >BMP, like this: >Diacritics are assigned to groups according to their position in the >diacritics sorting string. Mnemonic ideas are great for your national standard. However, it is too late for the international layout for the Latin script. But in looking at it, it is relatively respecting this idea already, I had never thought about it. >Glyps are produced by combining level and basic grapheme. >Thus, if caron is group n, k-caron (UXXXX) is typed like this: "GROUP N >SELECTING KEY (COMBINATION)" + "K". >Gaps are ignored, or since n-caron is not part of BMP, the same procedure >would give nothing for ... + "N" (or, the system could be without limits, >so that every nonexisting glyph image is produced by the combine option). > >Logically, the cyrillic users would like the same. Generally speaking, when >it comes to cyrillic, the standardisation situation is chaotic. One >keyboard standard among native users is there (although I do not know for >non-slavic glyphs), and MANY asdf-type cyrillic standards in the west (the >ja-sh-e-r-t-y things). This (a K3 issue) should definitely be sorted out. >Since cyrillic has more independent letters and less diacritics a >multidimensional representation (K2 again) is probably harder (though one >might try, and even cover over half of the cyrillic glyphs of BMP without >being left to arbitrary assigning of positions). > > >K3. >- A national ekyboard, with a set of groups, for (say) Icelandic, Danish, >US, or (say) Northern Sámi, Finnish and English. > > > >I was in London and talked to Yves, it was very interesting, and he >promised to send me some stuff. We talked about a fourth level, at this >point I am afraid he did not quite understand me. My main concern was not >Sámi at that discussion, but the possibility of typing the glyphs of the >entire code page, and of finding them in a position easy to remember. > >His main concern is the key placement issue (rather than the assignment of >glyphs to keys), needless to say, there still are many overlaping cases, >and I am looking forward to reading his things. > >___________________________________________________________________ >Trond Trosterud email: trondt@barsek.hsf.no >Barentssekretariatet, P.O.Box 276, work: +47-7899-3758 >N-9901 Kirkenes, Norway fax: +47-7899-3225 >http://www.norut.no/barsek/ip/iphome.html home: +47-7899-2243 >------------------------------------------------------------------- >Character test string, please ignore: á½¼†ªº‡-Á¢ƒ±¥µ…-âš›—˜-™¸œ–¯ >___________________________________________________________________ Alain LaBonté Québec