Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Path: utzoo!dciem!nrcaer!cognos!roberts From: roberts@cognos.uucp (Robert Stanley) Newsgroups: can.francais,can.general Subject: Re: Accents et objectifs Message-ID: <1202@cognos.UUCP> Date: Wed, 29-Jul-87 20:10:05 EDT Article-I.D.: cognos.1202 Posted: Wed Jul 29 20:10:05 1987 Date-Received: Sat, 1-Aug-87 11:46:32 EDT References: <424@pembina.UUCP> <228@Mannix.iros1.UUCP> <232@Mannix.iros1.UUCP> Reply-To: roberts@cognos.UUCP (Robert Stanley) Distribution: can Organization: Cognos Inc., Ottawa, Canada Lines: 58 Xref: dciem can.francais:28 can.general:661 Summary: back to EBCDIC In article <232@Mannix.iros1.UUCP> fortin@iros1.UUCP (Denis Fortin) writes: >Evidemment, je ne suggere pas de faire de l'ASCII-bang-bang un >standard pan-canadien (eeeeeek!) mais tout de meme, c'est mieux >avec des accents! Bien sur, c'est mieux avec des accents. But it's ever so much easier to enter data without accents, even when you have a nice dead-key system such as that used on the Macintosh. And I, for one, find reading these ugly inserted special characters a real impediment. What is more, they drive some of my regular expressions and smart search utilities crazy. Their only merit seems to be for display (read: print) purposes, unless some enterprising net person were to write a neat filter to translate (in both directions) between an agreed inserted accent standard and some agreed term-cap extension for terminals capable of displaying accented letters. However, I think this brings us round to the need for a character set which includes all possible accented letter combinations as unique codes, or a standard for encoding accents in text strings. An earlier posting suggested that 144 character codes would be required to enable a full complement of accented letters to be encoded as single characters. Unfortunately 144 into 128 doesn't go, and I have this strange feeling that 8-bit ASCII has already been grabbed by the graphics fraternity. Perhaps we should turn back to Big Blue and adopt EBCDIC as the universal (let us not be provincial in this matter) standard. To be serious, this is a problem that is going to have to be resolved, much as the Japanese are coming to terms with encoding and displaying their character sets. For all languages that require addition/alteration to the standard Roman alphabet, left to right entered and unaugmented, we need an internal representation which takes collating sequences into account, an agreed display standard, and an acceptable data-entry mechanism. Clearly there are a variety of techniques that can be employed to assist translate inbound and outbound characters, ranging from phonetic keyboards (Boswell, where are you?) to reasoning mechanisms capable of deducing accents that need to be displayed. But at the heart we need not just a consensus, but a standard. If there is this much interest in can.francais, perhaps a serious effort can be mounted, which might look (say) at all roman alphabet-based written languages, with a view to developing the step beyond ASCII. I wonder what happens to net traffic volumes when the basic character requires 14 bits! :-) Never let us forget that computers are tools, and we the users. We should not have to compromise our working standards simply because the tools are inadequate. If accents are important, which they certainly are, then let us work towards tools that makes it easy to employ them. After all, memory finally is cheap, data storage is getting cheap, and I suspect that high-speed data communications will be cheap in a year or three. << you better be prepared to make the most of the future, because that's where you're going to live the rest of your life>> -- Robert Stanley Compuserve: 76174,3024 Cognos Incorporated uucp: decvax!utzoo!dciem!nrcaer!cognos!roberts 3755 Riverside Drive or ...nrcaer!uottawa!robs Ottawa, Ontario Voice: (613) 738-1440 - Tuesdays only (don't ask) CANADA K1G 3N3