3 ms·

My favorite quote: About this method he said, “You don’t need a hydrology course to learn to swim. You don’t point at the water and say, ‘This is water, this i

My favorite quote: About this method he said, “You don’t need a hydrology course to learn to swim. You don’t point at the water and say, ‘This is water, this is how water works.’ you just throw the babies in.”

peterburkimsher 9y ago
I agree completely. Language is alive, and learning words is much easier for me if they're in context. I'm not learning Latin, but I am learning Chinese. I wrote http://pingtype.github.io http://pingtype.github.io to put spaces between words, colours for tones, pinyin romanisation, and a literal translation on the line below. I could probably do the same for Latin. First I'm doing English-to-Chinese (the other way), then I might try Taiwanese and Japanese, but the structure of the program is all the same. I just have to gather the data for each new language. Comment if you're interested, and I'll let you know when I finally get around to working on it. There have been a few Latin/Greek posts on HN that get upvoted a lot, so I think there's some demand.
- mncharity 9y ago
  Nifty. I went looking for google translate style audio, and found the Advanced > Audio button, and hit a bug. Page load; for Chinese input, enter "中文輸入"; Translate; Audio; surprised by English input and Chinese spaced having cat and ffmpeg command lines. Feature idea: permit repeatedly playing the audio for just a part of the phrase. Often when using ggl translate, I want to rehear just a syllable or word I'm unsure of, or am having trouble with. Google makes you rehear the whole thing, which is simple, but discourages double-checks, and impairs iteration.
  - peterburkimsher 9y ago
    Sorry! Thank you for exploring the Advanced features, I put a lot of effort into those, but they need more explaining. I'm pleased that you actually looked at it though :). The audio feature doesn't actually play audio. It's supposed to generate input commands to put into ffmpeg. I would write more documentation, but my priority is to get the existing documentation translated to Chinese. I asked 4 friends over the past 3 months, all have promised to do it, but only one of them translated one page. I don't think there's a way to do what you're asking using only client-side JavaScript. I would need a server to generate an audio file for you. Unfortunately, I don't have the money for that. The most I could do would be to generate an M3U playlist file, which you could then save and navigate using the chapter feature of your music player.
    - mncharity 9y ago
      > get the existing documentation translated to Chinese http://opennmt.net/ http://opennmt.net/ links to a "Live system" with a Chinese option. Discussed on infoworld[1]. FWIW, its translation of "This is a test." was at least recognized by google translate. > audio [...] using only client-side JavaScript It looks like the audio starts out as small mp3 files, which are then concatenated? There are WebAudio-based libraries both for playing one file after another, and for concatenating them for download.[2] For the mp4 subtitles, hmm, well, even if nothing else, there's ffmpeg.js[3] (but fonts might be an issue?). Thanks for your work. :) [1] http://www.infoworld.com/article/3152050/artificial-intelligence/open-source-challenger-takes-on-google-translate.html http://www.infoworld.com/article/3152050/artificial-intellig... [2] For example (merely the first one I saw): https://github.com/jackedgson/crunker https://github.com/jackedgson/crunker [3] https://github.com/Kagami/ffmpeg.js https://github.com/Kagami/ffmpeg.js
- JoeDaDude 9y ago
  I'd be interested in a Latin version. Your description sounds exactly like what I need.
mncharity 9y ago
Thank you for doing the excerpt! I'd not have read the OP otherwise. And it turned out to provide additional perspective on describing a current project in science and engineering education transformation. Like Rome, the physical world has science and engineering stories that are highly connected, richly interwoven, and best taught that way. By exploring in conversation with one who is familiar with them. Like the "hydrology" example, hydrodynamics is best taught by swimming (in VR/AR) - sure there's a Reynolds number order-of-magnitude scale and slider on the side, and length and time, and a live Navier-Stokes on the wall, but that's all in service to playfully messing around in/with fluids. In K-4. Like Latin, the way science is taught doesn't work, doesn't inspire, and leaves out lots of accessible engaging stories. Like Latin, perhaps there's hope for change. Unlike the instruction, you really need an conversation system with eye tracking, to notice your interests, answer questions, and mentor understanding. An expert and teacher to walk with. Anyway, my thanks.

peterburkimsher 9y ago

I agree completely. Language is alive, and learning words is much easier for me if they're in context. I'm not learning Latin, but I am learning Chinese. I wrote http://pingtype.github.io http://pingtype.github.io to put spaces between words, colours for tones, pinyin romanisation, and a literal translation on the line below. I could probably do the same for Latin. First I'm doing English-to-Chinese (the other way), then I might try Taiwanese and Japanese, but the structure of the program is all the same. I just have to gather the data for each new language. Comment if you're interested, and I'll let you know when I finally get around to working on it. There have been a few Latin/Greek posts on HN that get upvoted a lot, so I think there's some demand.

mncharity 9y ago
Nifty. I went looking for google translate style audio, and found the Advanced > Audio button, and hit a bug. Page load; for Chinese input, enter "中文輸入"; Translate; Audio; surprised by English input and Chinese spaced having cat and ffmpeg command lines. Feature idea: permit repeatedly playing the audio for just a part of the phrase. Often when using ggl translate, I want to rehear just a syllable or word I'm unsure of, or am having trouble with. Google makes you rehear the whole thing, which is simple, but discourages double-checks, and impairs iteration.
- peterburkimsher 9y ago
  Sorry! Thank you for exploring the Advanced features, I put a lot of effort into those, but they need more explaining. I'm pleased that you actually looked at it though :). The audio feature doesn't actually play audio. It's supposed to generate input commands to put into ffmpeg. I would write more documentation, but my priority is to get the existing documentation translated to Chinese. I asked 4 friends over the past 3 months, all have promised to do it, but only one of them translated one page. I don't think there's a way to do what you're asking using only client-side JavaScript. I would need a server to generate an audio file for you. Unfortunately, I don't have the money for that. The most I could do would be to generate an M3U playlist file, which you could then save and navigate using the chapter feature of your music player.
  - mncharity 9y ago
    > get the existing documentation translated to Chinese http://opennmt.net/ http://opennmt.net/ links to a "Live system" with a Chinese option. Discussed on infoworld[1]. FWIW, its translation of "This is a test." was at least recognized by google translate. > audio [...] using only client-side JavaScript It looks like the audio starts out as small mp3 files, which are then concatenated? There are WebAudio-based libraries both for playing one file after another, and for concatenating them for download.[2] For the mp4 subtitles, hmm, well, even if nothing else, there's ffmpeg.js[3] (but fonts might be an issue?). Thanks for your work. :) [1] http://www.infoworld.com/article/3152050/artificial-intelligence/open-source-challenger-takes-on-google-translate.html http://www.infoworld.com/article/3152050/artificial-intellig... [2] For example (merely the first one I saw): https://github.com/jackedgson/crunker https://github.com/jackedgson/crunker [3] https://github.com/Kagami/ffmpeg.js https://github.com/Kagami/ffmpeg.js
JoeDaDude 9y ago
I'd be interested in a Latin version. Your description sounds exactly like what I need.

mncharity 9y ago

Thank you for doing the excerpt! I'd not have read the OP otherwise. And it turned out to provide additional perspective on describing a current project in science and engineering education transformation. Like Rome, the physical world has science and engineering stories that are highly connected, richly interwoven, and best taught that way. By exploring in conversation with one who is familiar with them. Like the "hydrology" example, hydrodynamics is best taught by swimming (in VR/AR) - sure there's a Reynolds number order-of-magnitude scale and slider on the side, and length and time, and a live Navier-Stokes on the wall, but that's all in service to playfully messing around in/with fluids. In K-4. Like Latin, the way science is taught doesn't work, doesn't inspire, and leaves out lots of accessible engaging stories. Like Latin, perhaps there's hope for change. Unlike the instruction, you really need an conversation system with eye tracking, to notice your interests, answer questions, and mentor understanding. An expert and teacher to walk with. Anyway, my thanks.