doc/chatlog.txt (f97c57b) - fs-workbench.git

chatlog.txt

(18:36:48) RayZa1: fenris_kcf: We were thinking about having a transformation DB right?
(18:37:04) fenris_kcf: öhm
(18:37:16) fenris_kcf: you mean for the tone-transitions?
(18:37:20) RayZa1: yes
(18:37:28) RayZa1: sound-transitions
(18:37:31) fenris_kcf: yes, sry
(18:37:35) fenris_kcf: i could create one
(18:37:43) fenris_kcf: atm there is no such DB
(18:38:46) RayZa1: I guess it might be helpful. We could create all sorts of visual representations like some sort of node diagram
(18:39:21) RayZa1: so you see on one page which transitions have been applied to which source and what the result is
(18:42:37) RayZa1: you could also have an algorithm to apply transformations to the word in the source language in order to achieve the specified word in the resulting language
(18:44:08) RayZa1: for "braudana" in PG and "Brot" in German that would be -"ana", "au">"o", "d">"t"
(18:45:13) RayZa1: for e.g. Danish "brød" it would be -"ana", "au">"ø"
(18:45:31) fenris_kcf: i see
(18:45:33) RayZa1: in case you'd look at direct transitions
(18:46:24) RayZa1: you can think about having intermediate states "Old High German", "Old Norse" and what not
(18:48:57) RayZa1: additionally it would be good to have an etymological DB for cases where the correct translation wouldn't be etymologically connected
(18:53:27) RayZa1: we had a few false friends in terms of translations already which were however connected in etymology
(18:54:51) RayZa1: also another thing one might somehow want to save is: "sprechen" and "talar" aren't connected etymologically but "talar" and "språk" are logically, and "Sprache" and "språk" are etymologically connected
(18:59:31) fenris_kcf: i agree that it would be helpful, but it will take some time to create such a system
(19:03:31) RayZa1: just to finish my thought: to come back to the transitions: one could then analyze those and decide which should be considered for FS in general. work on multiple representative examples to get a representative result and then have the transitions applied automatically. should you feel you made the wrong assumptions in general you can easily correct the transitions being applied or if you feel the algorithms works well in general bu
(19:04:05) RayZa1: *algorithm
(19:04:49) RayZa1: yes I agree such a system is not trivial but it's what's running in your head anyway
(19:05:26) RayZa1: and this way everyone can see the "reasoning" without having to explain everything in length
(19:06:09) RayZa1: you can also more easily try out things
(19:06:59) fenris_kcf: maybe we can develop such a system together
(19:07:10) RayZa1: that's what I was thinking as well
(19:07:47) fenris_kcf: let me create a repository ...
(19:09:25) RayZa1: I have to tell you know, I know Java, C#, Python and JavaScript but I don't know PHP
(19:09:34) RayZa1: tell you though*
(19:11:08) fenris_kcf: then i'd say Javasript or python
(19:11:17) fenris_kcf: though i can program in Java, i don't likve to
(19:11:19) fenris_kcf: *like
(19:11:38) fenris_kcf: never programmed in C#
(19:11:40) RayZa1: I'm rusty in Java anyway
(19:11:58) RayZa1: we could try out how well it works with Node.js?
(19:12:13) RayZa1: I haven't really developed for Node.js yet but I wouldn't mind
(19:12:33) fenris_kcf: since the system will probably be used in the web, i vote for JS
(19:12:43) fenris_kcf: Node-stuff can used on server side JS
(19:12:55) fenris_kcf: but at first i would avoid using Node.js
(19:13:16) RayZa1: yes it should work both on client and server side
(19:13:40) RayZa1: I really like classes though :S
(19:13:49) RayZa1: what to do?
(19:14:18) RayZa1: it's the thing that annoyed me the most with JS, that there was no easy way to do classes
(19:14:52) RayZa1: or no elegant way either
(19:15:07) fenris_kcf: you can use classes since ES6
(19:15:46) fenris_kcf: https://git.schokokeks.org/fs-workbench.git/
(19:15:53) RayZa1: but they won't work in all browsers yet
(19:15:56) fenris_kcf: you can use the public key, that you gave me
(19:16:05) RayZa1: ok
(19:16:13) fenris_kcf: that's not a problem, RayZa1, since one can convert ES6 to ES5
(19:16:18) fenris_kcf: using Babel for example
(19:17:37) fenris_kcf: https://babeljs.io/
(19:18:49) RayZa1: ok
(19:19:12) RayZa1: then ES6

fs-workbench.git

added small test