sutela@polaris.utu.fi (Kari Sutela) (10/31/90)
ifarqhar@sunc.mqcc.mq.oz.au (Ian Farquhar) writes: >In article <sutela.656751981@polaris> sutela@polaris.utu.fi (Kari Sutela) writes: [I wrote about hardcoding a language on a chip] >Do phonems vary greatly from language to language? I'm not a linguist, but I'd think that they do (at least, a bit). I, for example, have had difficulties in producing reasonably sounding Finnish speech. On the other hand, this might be a problem with all languages --- the phonemes are just an approximation of the real human ones (or perhaps I just haven't tried hard enough). Anyway, if you think about non-european languages (for example, some african languages), I'd guess that there are indeed phonemes which can't be produced with the ones which we are using. >One feature that I would like to see in AmigaOS is the provision of >support for internationalism ie. a library that knows what time zone >the user is in, what their system of measurement is (ie. metric or >otherwise), the language, the currency, date formats etc. Exactly, I agree. On the other hand, I seem to remember a vague comment (by someone from CATS?) about a locale.library in OS 2.0 --- is there such a beast in 2.0? I'd think that such a library should also provide functions for deciding word-breaking characters, etc. I find it annoying when a text-editor word-wraps and breaks a Finnish word into two, just because it considered an umlaut-a a non-word characte. Preferably all accented characters should be considered as word-characters --- one doesn't always write in one language, for example, foreign names could include strange, accented characters. Kari Sutela sutela@polaris.utu.fi -- Kari Sutela sutela@polaris.utu.fi
peterk@cbmger.UUCP (Peter Kittel GERMANY) (10/31/90)
In article <sutela.657357692@polaris> sutela@polaris.utu.fi (Kari Sutela) writes: >I'd think that such a library should also provide >functions for deciding word-breaking characters, etc. I find it annoying >when a text-editor word-wraps and breaks a Finnish word into two, just >because it considered an umlaut-a a non-word characte. Preferably >all accented characters should be considered as word-characters --- one >doesn't always write in one language, for example, foreign names could >include strange, accented characters. Gould point, why didn't this come from me? So here comes a PLEA TO ALL PROGRAMMERS: If you write a program that is considered to process text, precisely plain ASCII text, then use EVERY character that is available. (Well, a real control char still should remain a control char.) Two examples of how not to do it: 1. Type command in DOS with the hex option. Every char with the MSB set is displayed as a dot, even when it is quite a normal, readable char for other people (seems like I should write another enhancement request). 2. The less program used on earlier Fish disks as a replacement for more to look into the readmes. It just canceled the MSB, thus changing the 8-bit chars to some quite different 7-bit ones. And this in a time where Fred is so polite to let programmers put also their own language readmes (besides the English version) on the disk. They simply get crippled and look unprofessional. -- Best regards, Dr. Peter Kittel // E-Mail to \\ Only my personal opinions... Commodore Frankfurt, Germany \X/ {uunet|pyramid|rutgers}!cbmvax!cbmger!peterk
zerkle@iris.ucdavis.edu (Dan Zerkle) (11/01/90)
In article <sutela.657357692@polaris> sutela@polaris.utu.fi (Kari Sutela) writes: >ifarqhar@sunc.mqcc.mq.oz.au (Ian Farquhar) writes: > >>In article <sutela.656751981@polaris> sutela@polaris.utu.fi (Kari Sutela) writes: >[I wrote about hardcoding a language on a chip] > >>Do phonems vary greatly from language to language? > >I'm not a linguist, but I'd think that they do (at least, a bit). I, for >example, have had difficulties in producing reasonably sounding Finnish >speech. On the other hand, this might be a problem with all languages --- This is a really big problem. My former workplace has mostly developed and English text-to-speech interpreter (available RSN from Panasonic!). However, there was also a large amount of work done to get the thing to speak Japanese and Chinese. This is not simply a matter of getting it to understand the input. Rather, a tremendous amount of the work was "tuning" the various sounds to be more comprehensible and natural. As a side note, we started off from the MITtalk system, which in turn was developed from a system developed by the late great Dennis Klatt. After I showed off my Amiga and they listened to the voice, the folks at work were very sure that the Amiga text-to-speech was a derivative of one of these. So, if you hear a box with Panasonic on the outside talking in a manner similar to your Amiga (only much, much better), you'll know why they sound alike. -Dan Dan Zerkle zerkle@iris.ucdavis.edu (916) 754-0240 Amiga... Because life is too short for boring computers.