xai: Difference between revisions

From Lojban
Jump to navigation Jump to search
mNo edit summary
 
mNo edit summary
Line 1: Line 1:


I guess you think of something like "Japanese Language Kit (JLK)" for Mac. This is an ''extension'' parallel to Macintosh "Chinese Language Kit (CLK)" what I'm using. Since I don't intend to write Japanese (Kana+Kanji) on my machine, this solution is far too expensive and sophisticated for the purpose of just typing in some Lojban-Hiragana syllables.
'''xai'''     KOhA    they. Repeats two or more preceding sumti, as ''ra'' repeats one. (proposed by [[jbocre: Pierre Abbat|Pierre Abbat]])


Yet, there could be still another way: recently, I downloaded some other fonts together with a software allowing to create kind of scripts in order to access every place on your keyboard you want, e.g. "a" -> "xy". I just don't know yet how to have that little program installed... -.aulun.
* [[User:And Rosta|And Rosta]] think this doesn't merit a precious monosyllablc cmavo: cp ''[[jbocre: ru'ai|ru'ai]]''


----
-----


''Getting dangerously off-topic here, but the language kits come free with System 9. Lamentably, MacOS X is yet to get its sh*t together wth customisable (or not!) Unicode keyboards, though at least it does allow you to use Unicode fonts. (System 9 only claims to.) For people who might care, could someone say whether there are shareware/freeware PC and Unix input methods? Or point to a good website? -- [[User:Nick Nicholas|nitcion]]''
* le gerku cu jersi le mlatu .i xai tatpi
** ''This is actually no less vague than "le gerku cu jersi le mlatu .i tatpi"''


* There is a Unicode text editor for UNIX (maybe even available for Windows) called [http://www.yudit.org udit]. In it you can define arbitrary key maps producing any [[jbocre: Unicode|Unicode]] character. (I think it is limited to plane 0, though.)Getting Unicode characters working everywhere else in Unix (or X) is significantly more difficult. See the [http://www.linuxdoc.org/HOWTO/Unicode-HOWTO.html nicode HOWTO] for Linux, 90% of which is applicable to every form of Unix.
***''le gerku'' and ''le mlatu'' are the '''only''' preceding sumti, so in this case ''xai'' is very precise: it can only be ''le gerku kujoi le mlatu''. On the other hand, a plain ''tatpi'' could refer to just ''le gerku'', just ''le mlatu'' or to neither.
**** Why {joi} rather than just {e}?  It seems each of them is tired, not the mass of them (can a mass be tired?). pc


----
***** I picked ''joi'' because I think it's the most useful. To get "each of them" we can say ''ro xai'', and to emphasize explicitly the mass meaning we can say ''piro xai''. In any case, whatever applies to ''do'', ''ta'' and company should apply to ''xai''. The Book is not very definite about this.
***** Well it's certainly going to have to be worked out whether "xai" means ".e" or "joi". The two sumti are not connected with any such connector in the first sentence, so that has to be created in the second. Also, does xai import ''every'' sumti from the previous sentence, and glom them together using as-yet-unspecified connector? What is meant by "mi klama abu. by. .i xai broda"?


Don't think that this is off-topic, but useful for further efforts with creating alternative Lojban scripts (and their specific input!).
****** It does not import every sumti from the previous sentence, nor do all the sumti it imports need be in the immediately preceding sentence. The proposed definition says "two or more preceding sumti" and I think it is a good one. ''ra'' is similarly vague as to which sumti it sends back to. As for ''mi klama abu. by. .i xai broda'', first notice that ''abu by'' is a single sumti, probably ''abuboi by'' is what was meant. In that case, the most likely interpretation is that ''xai'' selects ''abu'' and ''by'', because talking about two locations together is more likely than talking about a goer and a location. Compare with ''John went from Paris to Rome. They are beautiful cities." It is unlikely that John is meant to be a part of "they". It depends on the context. --[[User:xorxes|xorxes]]
******* This is absolutely NO better than dropping the sumti completely and letting an implied zo'e represent all that.


Since always having been working with Mac (and Atari) machines, I know next to nothing about .uindoz. stuff. But there might be a solution:
********"two or more preceding sumti" is rather more informative than ''zo'e''. Would you also maintain that using ''ra'' is no better than dropping the sumti completely and letting an implied zo'e represent that?
********* ra is clearer than xai and zo'e, because it specifies the number of sumties to copy: one. xai and zo'e both leave the number, and the relationship (.e, joi, etc) undetermined, so they are totally equal.


Contact SIL (Summer Institute for Linguistics) in Dallas/Tx and ask for shareware "SILKey" (a package of the program SILKey, SIL extension, SIL Weaver and some sample files like Pig Latin kbd and a tutorial). This is for Macintosh. But there's also hope for DOSen-user ;-), since there's "excellent" KeyMan for Windows! ('''''www.sil.org''''').
********* ''xai'' specifies at least two sumti, this often will provide more information than we can get for ''ra''. ''zo'e'' does not even have to be about any previous sumti. Very often it is not.
********** ra means ''one'' sumti; xai means anywhere between 2 and ci'i, so you can see that ra is more well-defined/better specified/less vague/more specific. Likewise, it is true that xai is better specified than zo'e.


Using the ''weaver'' software, one is able to generate kbd-files in order to manipulate the keyboard access in many different ways... (just gave it a rudimentary try - yet it works!) - .aulun.
********** ''ra'' normally has fewer possible referents, that's true. (The only exception I can think of is when there are exactly two preceding sumti, in which case both ''xai'' and ''ra'' are perfectly well defined.) But that doesn't make it less vague. In a situation for example where we have seven preceding sumti, two of which are obviously related and form a natural group, and the rest are unrelated, ''xai'' has a much more likely obvious referent than ''ra''.


----
**I take back what I said elsewhere about this problem having an easy solution if we just invent new words.  {xai} needs for success at least 1) a way of indicating which recent sumti are intended, unless it is all back to the beginning of the conversation, and 2) a way to indicate how they are to be connected, minimally to distinguish {e} and {joi} -- maybe two separate pronouns or taking one as standard and using something from LAhE to get the other.  And, for 1, maybe a complex way of indicating which is involved -- backcounting for example.  It seems that simple repetition is likely to be easier and clearer.
***For absolute precision you can use subscripts. Just as ''raxipaci'' represents the 13th sumti back, ''xaixipacipi'epabipi'ecimu'' can be used to refer to the 13th, 18th and 25th previous sumti taken together.


Here's a keyboard script that I wrote (within "Weaver" - prior to compilation and installation). The script addresses the original font's places '''except''' for the consonant-n's and the (new) L-row's places (of my own modified font-version):
****don't you mean rixipaci?  ra is ambiguous.  Furthermore, anyone who uses a subscript to refer back to the 13th sumti should be shot (or more nicely said: will not be understood). An ambiguous xai would be *far* more useful than any attempt to make it specific (if you need the specificity, there's nothing wrong with going back to goi ko'a and goi ko'e, ko'a joi ko'e). -- mi'e .djorden.
 
*****Of course the plain ''xai'' is the useful one (and the only usable one) but nothing is lost by providing a method for the precision that some people seem to require.
// Hiragana keyboard AWT 8.01
 
// access to syllabic chars
 
keyboardname "U.S."
 
initialgroup (Main)
 
store(vowel) "aiueo"
 
store(hiravow) "qwert"
 
store(marks) ".,;"
 
store(hiramrk) '?"'
 
group (Main) using keys
 
"t"+"a" > "y"
 
"t"+"i" > "u"
 
"ch"+"i" > "u"
 
"ts"+"u" > "i"
 
"t"+"u" > "i"
 
"t"+"e" > "o"
 
"t"+"o" > "p"
 
"y"+"a" > "Q"
 
"y"+"i" > "W"
 
"y"+"u" > "E"
 
"y"+"e" > "R"
 
"y"+"o" > "T"
 
"m"+"a" > "Y"
 
"m"+"i" > "U"
 
"m"+"u" > "I"
 
"m"+"e" > "O"
 
"m"+"o" > "P"
 
"k"+"a" > "a"
 
"k"+"i" > "s"
 
"k"+"u" > "d"
 
"k"+"e" > "f"
 
"k"+"o" > "g"
 
"n"+"a" > "h"
 
"n"+"i" > "j"
 
"n"+"u" > "k"
 
"n"+"e" > "l"
 
"n"+"o" > ";"
 
"w"+"a" > "A"
 
"w"+"i" > "S"
 
"w"+"u" > "D"
 
"w"+"e" > "F"
 
"w"+"o" > "G"
 
"v"+"a" > "A"
 
"v"+"i" > "S"
 
"v"+"u" > "D"
 
"v"+"e" > "F"
 
"v"+"o" > "G"
 
"r"+"a" > "H"
 
"r"+"i" > "J"
 
"r"+"u" > "K"
 
"r"+"e" > "L"
 
"r"+"o" > ":"
 
"l"+"a" > "�" // produces special chars for L-syllables (altered by myself and not included in the font available)...
 
"l"+"i" > "�"
 
"l"+"u" > "�"
 
"l"+"e" > "�"
 
"l"+"o" > "�"
 
"s"+"a" > "z"
 
"s"+"i" > "x"
 
"sh"+"i" > "x"
 
"s"+"u" > "c"
 
"s"+"e" > "v"
 
"s"+"o" > "b"
 
"h"+"a" > "n"
 
"h"+"i" > "m"
 
"h"+"u" > ","
 
"f"+"u" > ","
 
"h"+"e" > "."
 
"h"+"o" > "/"
 
"d"+"a" > "�"
 
"d"+"i" > "$"
 
"j"+"i" > "$"
 
"dj"+"i" > "$"
 
"d"+"u" > "%"
 
"dz"+"u" > "%"
 
"d"+"e" > "�"
 
"d"+"o" > "1"
 
"g"+"a" > "�"
 
"g"+"i" > "�"
 
"g"+"u" > "�"
 
"g"+"e" > "�"
 
"g"+"o" > "�"
 
"z"+"a" > "*"
 
"z"+"i" > "�"
 
"z"+"u" > "V"
 
"z"+"e" > "�"
 
"z"+"o" > "�"
 
"b"+"a" > "!"
 
"b"+"i" > "�"
 
"b"+"u" > "2"
 
"b"+"e" > "3"
 
"b"+"o" > "�"
 
"p"+"a" > "�"
 
"p"+"i" > "�"
 
"p"+"u" > "�"
 
"p"+"e" > "�"
 
"p"+"o" > "�"
 
"k"+"l" > "[jbocre: "  //Some still provisory commands to produce some small-size chars ...
 
"k"+"k" > "]"
 
"k"+"j" > "{"
 
"k"+"m" > "}"
 
"n"+"n" > "�" //hitting the n-key '''twice''' produces Japanese consonant-n
 
any(vowel) > index(hiravow,1) fix
 
any(marks) > index(hiramrk,1) fix
 
----

Revision as of 17:20, 4 November 2013

xai KOhA they. Repeats two or more preceding sumti, as ra repeats one. (proposed by Pierre Abbat)

  • And Rosta think this doesn't merit a precious monosyllablc cmavo: cp ru'ai

  • le gerku cu jersi le mlatu .i xai tatpi
    • This is actually no less vague than "le gerku cu jersi le mlatu .i tatpi"
      • le gerku and le mlatu are the only preceding sumti, so in this case xai is very precise: it can only be le gerku kujoi le mlatu. On the other hand, a plain tatpi could refer to just le gerku, just le mlatu or to neither.
        • Why {joi} rather than just {e}? It seems each of them is tired, not the mass of them (can a mass be tired?). pc
          • I picked joi because I think it's the most useful. To get "each of them" we can say ro xai, and to emphasize explicitly the mass meaning we can say piro xai. In any case, whatever applies to do, ta and company should apply to xai. The Book is not very definite about this.
          • Well it's certainly going to have to be worked out whether "xai" means ".e" or "joi". The two sumti are not connected with any such connector in the first sentence, so that has to be created in the second. Also, does xai import every sumti from the previous sentence, and glom them together using as-yet-unspecified connector? What is meant by "mi klama abu. by. .i xai broda"?
            • It does not import every sumti from the previous sentence, nor do all the sumti it imports need be in the immediately preceding sentence. The proposed definition says "two or more preceding sumti" and I think it is a good one. ra is similarly vague as to which sumti it sends back to. As for mi klama abu. by. .i xai broda, first notice that abu by is a single sumti, probably abuboi by is what was meant. In that case, the most likely interpretation is that xai selects abu and by, because talking about two locations together is more likely than talking about a goer and a location. Compare with John went from Paris to Rome. They are beautiful cities." It is unlikely that John is meant to be a part of "they". It depends on the context. --xorxes
              • This is absolutely NO better than dropping the sumti completely and letting an implied zo'e represent all that.
                • "two or more preceding sumti" is rather more informative than zo'e. Would you also maintain that using ra is no better than dropping the sumti completely and letting an implied zo'e represent that?
                  • ra is clearer than xai and zo'e, because it specifies the number of sumties to copy: one. xai and zo'e both leave the number, and the relationship (.e, joi, etc) undetermined, so they are totally equal.
                  • xai specifies at least two sumti, this often will provide more information than we can get for ra. zo'e does not even have to be about any previous sumti. Very often it is not.
                    • ra means one sumti; xai means anywhere between 2 and ci'i, so you can see that ra is more well-defined/better specified/less vague/more specific. Likewise, it is true that xai is better specified than zo'e.
                    • ra normally has fewer possible referents, that's true. (The only exception I can think of is when there are exactly two preceding sumti, in which case both xai and ra are perfectly well defined.) But that doesn't make it less vague. In a situation for example where we have seven preceding sumti, two of which are obviously related and form a natural group, and the rest are unrelated, xai has a much more likely obvious referent than ra.
    • I take back what I said elsewhere about this problem having an easy solution if we just invent new words. {xai} needs for success at least 1) a way of indicating which recent sumti are intended, unless it is all back to the beginning of the conversation, and 2) a way to indicate how they are to be connected, minimally to distinguish {e} and {joi} -- maybe two separate pronouns or taking one as standard and using something from LAhE to get the other. And, for 1, maybe a complex way of indicating which is involved -- backcounting for example. It seems that simple repetition is likely to be easier and clearer.
      • For absolute precision you can use subscripts. Just as raxipaci represents the 13th sumti back, xaixipacipi'epabipi'ecimu can be used to refer to the 13th, 18th and 25th previous sumti taken together.
        • don't you mean rixipaci? ra is ambiguous. Furthermore, anyone who uses a subscript to refer back to the 13th sumti should be shot (or more nicely said: will not be understood). An ambiguous xai would be *far* more useful than any attempt to make it specific (if you need the specificity, there's nothing wrong with going back to goi ko'a and goi ko'e, ko'a joi ko'e). -- mi'e .djorden.
          • Of course the plain xai is the useful one (and the only usable one) but nothing is lost by providing a method for the precision that some people seem to require.