Post by Auburn on Dec 23, 2014 3:04:17 GMT
From Skype, for reference.
Fifi and I talking about how to implement word splitting in languages such as French.
[6:12:17 PM] Dreamingfifi: by the way, what would you do about circumfixes and infixes?
[6:13:17 PM] Dreamingfifi: a circumfix is something like in French: ne ... pas, where the negation is put around the verb
[6:14:04 PM] Dreamingfifi: an infix is put inside a verb.. i believe that is how roots are conjugated in Arabic, as well as in Adûnaic
[6:15:51 PM] Auburn II: humm
[6:18:53 PM] Auburn II: gosh, it's hard to know what to do with some languages. yknow I was imagining that if this software was successful it would be a sort of version 1.0 - with certain features. and if it does well enough then a more sophisticated interface with more abilities can be generated.
I just wouldn't want to have it be something that takes a programmer months to write xD cuz they'd charge a $7,000 bill or something. But on the other hand, French isn't some obscure language. It's one of the more common ones so maybe we should account for it.
[6:19:40 PM] Dreamingfifi: I actually have an idea for how to handle the problem of infixes and circumfixes
[6:19:45 PM] Auburn II: yea?
[6:20:13 PM] Dreamingfifi: allow for split words
[6:21:42 PM] Dreamingfifi: so... like this. you have words in order
1
2
3
4
[end sentence]
allow for:
1
2
3.1
4
3.2
end sentence
[6:22:25 PM] Dreamingfifi: both 3.1 and 3.2 would activate the same word on the next column, and have the same notations
[6:25:03 PM] Auburn II: hehe. that works!
[6:25:48 PM] Dreamingfifi: this could work another way, say that instead of the columns Sindarin - English - Notes you have Sindarin - French - Notes
then you'd have the word worder like this (numbers indicating not te order but the words)
Sindarin:
1 2 3 4.
French:
1 2.1 3 2.2 4.
[6:28:05 PM] Dreamingfifi: Then rolling over the sentence, the split word would show up like this:
Legolas [ú-]vêl Arwen.
Legolas [n']aime [pas] Arwen.
[6:28:26 PM] Dreamingfifi: brackets showing high-lighting
[6:28:37 PM] Auburn II: ahaa
[6:29:44 PM] Dreamingfifi: you'd have to allow then to re-order the words for each collum as needed though
[6:30:06 PM] Auburn II: I was just thinking about that; the columns. like the entry page...
[6:30:58 PM] Dreamingfifi: maybe a [slipt entry into 2] then next to each a little ticker which slets you set the word-order
[6:31:13 PM] Dreamingfifi: *ticker>menu
[6:32:54 PM] Auburn II: right. and how about -- in order to keep the columns mostly unchanged -- we give things numbers like:
Sindarin:
1
2
3
4
French
1
2
3.5
3
4
---- where 2 and 3.5 are actually the same sindarin word split into two. but the order of display will be correct: 1, 2, 3, 3.5, 4.
[6:33:17 PM] Auburn II: but the software will remember that 2 and 3.5 are connected, so the highlighter will correspond to both
[6:33:44 PM] Auburn II: *order of display on the final outcome page
[6:35:07 PM] Dreamingfifi: in my thing, the numbers were not for order, but for the database to keep track of the entries
[6:36:05 PM] Dreamingfifi: and then the numbers for ordering would look the same...
Sindarin : 1 2 3 4
French: 1 2 3 4 5
[6:36:08 PM] Auburn II: yea. the connection between the highlights right? ...but I think that can be tracked as a background process of the programming. The thing that needs user customization is word order of appearance on the display screen, I think
[6:37:04 PM] Auburn II: -still thinking though- it's tricky to visualize all this
[6:37:40 PM] Dreamingfifi: yeah, and based on how many entries, counting the split entires as individual words to be ordered with simple prime numbers
[6:38:59 PM] Dreamingfifi: so, the ú- entry would looke like this:
[2] [ ú-] = [2] [ n'] [4] [ pas ] grammatical stuff
[6:39:46 PM] Dreamingfifi: 2= second word in the sentence
4= 4th word in the sentence
[6:43:09 PM] Dreamingfifi: oh! Idea!!! At the bottom of the page, you'd be able to see the resulting columns to make sure you'd getting it right as you work!
[6:43:22 PM | Edited 6:45:53 PM] Auburn II: ah okay, I see. that works too. if i'm not mistaken, this eliminates the need for the decimal splits, right? ...cause, y'see I was thinking that even if an entry was "split", it would just expand the row where the sindarin word is.
so if it was ú-, then it be:
1: Legolas = Legolas [order:1]
2: ú- = 'n [order: 2]
- - - - - pas [order: 4]
3: vêl = aime [order: 3]
4: Arwen = Arwen [order: 5]
[6:43:40 PM] Auburn II: mm? lemmy think about that
[6:44:47 PM] Auburn II: er, to finish that thought ^ -- so then the row for "2" would contain both 2 and 4, and the program would know this, and connect the two via a background process. but the display order of both of those parts ('n and pas) would appear as one would want them to.
[6:49:27 PM] Dreamingfifi: so, the whole phrase would be:
[1][Legolas] = [1][Legolas]
[2][ú-] = [2][n'] [4][pas]
[3][vêl] = [3][aime]
[4][Arwen] = [5][Arwen]
[6:49:27 PM] Auburn II: yeah
[6:52:28 PM] Dreamingfifi: then going the other way:
French->English
[1] [Legolas] = [1][Legolas]
[2][n'] [4][pas] = [2][doesn't]
[3][aime] = [3][like]
[5][Arwen] = [4][Arwen]
[6:54:18 PM] Dreamingfifi: so in the database, each entry would have a code, which the user wouldn't see, and would connect the words together in the columns and makes the proper notation to show up
[6:55:42 PM] Auburn II: yeah, exactly. the less things the user needs to keep track of the more intuitive we can make it. =)
[6:55:56 PM] Dreamingfifi: yup
[6:56:22 PM] Auburn II: awesome! we should copypaste this conversation to some thread. lol.
[6:56:45 PM] Dreamingfifi: in longer sentences, i could see a dropdown menu that lets you select the word order getting clumsy
[6:57:05 PM] Dreamingfifi: which is why some sort of live-preview would be handy
[6:58:10 PM] Auburn II: yes. it definitely would. i'll ask the programmer how hard it'd be to add a previewer. if it's doable I think that'd be great.
Fifi and I talking about how to implement word splitting in languages such as French.
[6:12:17 PM] Dreamingfifi: by the way, what would you do about circumfixes and infixes?
[6:13:17 PM] Dreamingfifi: a circumfix is something like in French: ne ... pas, where the negation is put around the verb
[6:14:04 PM] Dreamingfifi: an infix is put inside a verb.. i believe that is how roots are conjugated in Arabic, as well as in Adûnaic
[6:15:51 PM] Auburn II: humm
[6:18:53 PM] Auburn II: gosh, it's hard to know what to do with some languages. yknow I was imagining that if this software was successful it would be a sort of version 1.0 - with certain features. and if it does well enough then a more sophisticated interface with more abilities can be generated.
I just wouldn't want to have it be something that takes a programmer months to write xD cuz they'd charge a $7,000 bill or something. But on the other hand, French isn't some obscure language. It's one of the more common ones so maybe we should account for it.
[6:19:40 PM] Dreamingfifi: I actually have an idea for how to handle the problem of infixes and circumfixes
[6:19:45 PM] Auburn II: yea?
[6:20:13 PM] Dreamingfifi: allow for split words
[6:21:42 PM] Dreamingfifi: so... like this. you have words in order
1
2
3
4
[end sentence]
allow for:
1
2
3.1
4
3.2
end sentence
[6:22:25 PM] Dreamingfifi: both 3.1 and 3.2 would activate the same word on the next column, and have the same notations
[6:25:03 PM] Auburn II: hehe. that works!
[6:25:48 PM] Dreamingfifi: this could work another way, say that instead of the columns Sindarin - English - Notes you have Sindarin - French - Notes
then you'd have the word worder like this (numbers indicating not te order but the words)
Sindarin:
1 2 3 4.
French:
1 2.1 3 2.2 4.
[6:28:05 PM] Dreamingfifi: Then rolling over the sentence, the split word would show up like this:
Legolas [ú-]vêl Arwen.
Legolas [n']aime [pas] Arwen.
[6:28:26 PM] Dreamingfifi: brackets showing high-lighting
[6:28:37 PM] Auburn II: ahaa
[6:29:44 PM] Dreamingfifi: you'd have to allow then to re-order the words for each collum as needed though
[6:30:06 PM] Auburn II: I was just thinking about that; the columns. like the entry page...
[6:30:58 PM] Dreamingfifi: maybe a [slipt entry into 2] then next to each a little ticker which slets you set the word-order
[6:31:13 PM] Dreamingfifi: *ticker>menu
[6:32:54 PM] Auburn II: right. and how about -- in order to keep the columns mostly unchanged -- we give things numbers like:
Sindarin:
1
2
3
4
French
1
2
3.5
3
4
---- where 2 and 3.5 are actually the same sindarin word split into two. but the order of display will be correct: 1, 2, 3, 3.5, 4.
[6:33:17 PM] Auburn II: but the software will remember that 2 and 3.5 are connected, so the highlighter will correspond to both
[6:33:44 PM] Auburn II: *order of display on the final outcome page
[6:35:07 PM] Dreamingfifi: in my thing, the numbers were not for order, but for the database to keep track of the entries
[6:36:05 PM] Dreamingfifi: and then the numbers for ordering would look the same...
Sindarin : 1 2 3 4
French: 1 2 3 4 5
[6:36:08 PM] Auburn II: yea. the connection between the highlights right? ...but I think that can be tracked as a background process of the programming. The thing that needs user customization is word order of appearance on the display screen, I think
[6:37:04 PM] Auburn II: -still thinking though- it's tricky to visualize all this
[6:37:40 PM] Dreamingfifi: yeah, and based on how many entries, counting the split entires as individual words to be ordered with simple prime numbers
[6:38:59 PM] Dreamingfifi: so, the ú- entry would looke like this:
[2] [ ú-] = [2] [ n'] [4] [ pas ] grammatical stuff
[6:39:46 PM] Dreamingfifi: 2= second word in the sentence
4= 4th word in the sentence
[6:43:09 PM] Dreamingfifi: oh! Idea!!! At the bottom of the page, you'd be able to see the resulting columns to make sure you'd getting it right as you work!
[6:43:22 PM | Edited 6:45:53 PM] Auburn II: ah okay, I see. that works too. if i'm not mistaken, this eliminates the need for the decimal splits, right? ...cause, y'see I was thinking that even if an entry was "split", it would just expand the row where the sindarin word is.
so if it was ú-, then it be:
1: Legolas = Legolas [order:1]
2: ú- = 'n [order: 2]
- - - - - pas [order: 4]
3: vêl = aime [order: 3]
4: Arwen = Arwen [order: 5]
[6:43:40 PM] Auburn II: mm? lemmy think about that
[6:44:47 PM] Auburn II: er, to finish that thought ^ -- so then the row for "2" would contain both 2 and 4, and the program would know this, and connect the two via a background process. but the display order of both of those parts ('n and pas) would appear as one would want them to.
[6:49:27 PM] Dreamingfifi: so, the whole phrase would be:
[1][Legolas] = [1][Legolas]
[2][ú-] = [2][n'] [4][pas]
[3][vêl] = [3][aime]
[4][Arwen] = [5][Arwen]
[6:49:27 PM] Auburn II: yeah
[6:52:28 PM] Dreamingfifi: then going the other way:
French->English
[1] [Legolas] = [1][Legolas]
[2][n'] [4][pas] = [2][doesn't]
[3][aime] = [3][like]
[5][Arwen] = [4][Arwen]
[6:54:18 PM] Dreamingfifi: so in the database, each entry would have a code, which the user wouldn't see, and would connect the words together in the columns and makes the proper notation to show up
[6:55:42 PM] Auburn II: yeah, exactly. the less things the user needs to keep track of the more intuitive we can make it. =)
[6:55:56 PM] Dreamingfifi: yup
[6:56:22 PM] Auburn II: awesome! we should copypaste this conversation to some thread. lol.
[6:56:45 PM] Dreamingfifi: in longer sentences, i could see a dropdown menu that lets you select the word order getting clumsy
[6:57:05 PM] Dreamingfifi: which is why some sort of live-preview would be handy
[6:58:10 PM] Auburn II: yes. it definitely would. i'll ask the programmer how hard it'd be to add a previewer. if it's doable I think that'd be great.