2 weeks to build a language app (a music-style language trainer)

[ryanheise]

OK, a question. What import options would people find most useful?

••• Open file on device •••
- Open file from phone's media library
- Transfer file from computer
- Download file from webpage

If I were to implement ONE of these, which one would be the most useful?

Open file on device.
Most of my apps open from the iTunes media library, but opening from iCloud files is nice too.

Excellent project, and I enjoy your language log experiments too!

Thanks for the encouragement! It makes me want to try even harder to achieve this impossible goal I had set for myself.

It was actually through those experiments that you referred to that I started experimenting with playing back clips at gradually increasing lengths, and I knew this was a really cool way to listen. I can't wait to convert this idea into something that I can do while out on my walks.

I’m an iPhone user and a Speater (similar to your project) user. Speater has access to any mp3 or video file I have transferred to my phone via iTunes. It’s very easy to open a file in Speater.

I will listen to file x normally using the iPhone music app and work with it intensively in Speater.

Importing a file via Dropbox or files seems to be duplicating processes? Wouldn’t that leave me with two copies of the same files in my phone?

That's what I suspected. I'd certainly like to make it as convenient as possible in every respect. I think I will still start with the file import option just during development so that I can get the app to a releasable stage quickly. At that point, Apple will let me start sharing the app with BETA testers, and from there I'll be able to actually try different things out and see how people like them.

For now, I can only do the Android internal release (although even that seems to be having delays as people who've tried clicking on the invite link are seeing a "Not found" error. It must take Google a few days for a new app to register in their systems and propagate to all of their servers.)

Since noone has actually been able to access it, I'll share here the current screenshot (in all its amazing glory):



Obviously this looks awful, but I have a plan and I'm sticking to it The big |X| is reserving the spot where I eventually want to display a waveform, and possibly text. As for the "A B C D" section labels, I have some ideas about how I want these to animate around when they get merged. Although really animations won't have any impact on the usefulness of the app, especially since I really intend for this app to be mostly used while the phone is in your pocket and you have your headphones plugged in. What I would like to maybe consider is for people who have smart watches, to display the transcript on the watch face, and allow the buttons on the watch to control going to the previous or next sentence. Or just with regular headphones that have the media controls, my intention is to be able to fully control the app through that so that the phone can stay in your pocket and you can enjoy your walk (while people stare at you strangely as you attempt to mimic the audio.)

[ryanheise]

Signing off for day 3. I now have a basic data model (subject to change) and I also managed to get the import feature working one day ahead of schedule, with local files, and also as a bonus, I have iCloud and GDrive working. Play/pause will come tomorrow (so as yet, the app is still 100% useless).

Also tomorrow, I'll also make a decision on the different training routines I would like to build into the app. I know for a start that I want to integrate the music-style one for memorisation, and I would also like to integrate the routine I used in my listening attention-span experiments.

The music-style routine is this:

A,B,AB
C,D,CD
AB,CD,ABCD
E,F,EF
G,H,GH
EF,GH,EFGH
ABCD,EFGH,ABCDEFGH

And the routine I used for extending my listening attention span was something similar to this:

A,B,C,D,E,F,G,H,
AB,CD,EF,GH,
ABCD,EFGH,
ABCDEFGH

These two routines have very different attributes and therefore different uses.

If I understand correctly, the back chaining routine would go like this:

A,
B,AB,
C,BC,ABC,
D,CD,BCD,ABCD,
etc.

(edit: after posting this, I realised there are different versions of back chaining. The one above is based on the approach linked in the first post. I believe Pimsleur does something different in that it may just begin on line 4 above)

I do not think the back chaining routine will work once the chain gets too long, but it can definitely work on a smaller scale if applied to individual sections, so I have to think carefully about how to integrate it maybe at a nested level. I also thought of a new sort of routine that is a sister of back chaining:

A,
B,AB,
C,ABC,
D,ABCD,
E,ABCDE,
F,ABCDEF,
G,ABCDEFG

It has the same limitations as normal back chaining, so it would have to be integrated only at a nested level.

Regarding the auto-play feature, there would be some options to have it repeat each item a set number of times before moving on, and another option which sounds strange, but which is to pause immediately after you play an item. You would need to hit the play button on the headphones or phone to get it to play again, or press the next button on the headphones or phone to go to the next item. This mode is what I need for trying to learn how to recite a story from memory since I would want to hear it once and then take my time to try to recite it.

One more important part of the plan is that if a translation is available, this can be read out using text-to-speech as part of the auto play feature. So you would hear <audio> <translation> <audio> <translation>. The exact behaviour will be configurable, but this feature is essential for a mode where you want to learn the meaning and have it all delivered through you're headphones while you're doing other things (like walking in my case).

[tangleweeds]

If I understand correctly, the back chaining routine would go like this:

A,
B,AB,
C,BC,ABC,
D,CD,BCD,ABCD,
etc.

(edit: after posting this, I realised there are different versions of back chaining. The one above is based on the approach linked in the first post. I believe Pimsleur does something different in that it may just begin on line 4 above).

The back chaining I have used, both in Japanese and in music, goes:

D
CD
BCD
ABCD

That way, the material gets more familiar the further you go in. The problem with backchaining from the front is that you're always moving from more familiar to less, always getting harder & requiring ever more brainpower, so you feel depleted much sooner and learn less. Doing it as above, you always start with the hardest bit, and it gets ever more familiar from there, so the exercise feels less depleting.

When you're memorizing a piece of music, the insider tip is *never* start learning from the beginning, always work from the end, for the reason above. And in performance, recording, performance-practice you'll playing it from the front on as per norm, so the backchaining and normal usage balance each other out. I'm not sure how frequently you're going to "perform" by listening to any one clip all the way though, though

[ryanheise]

The back chaining I have used, both in Japanese and in music, goes:

D
CD
BCD
ABCD

That way, the material gets more familiar the further you go in. The problem with backchaining from the front is that you're always moving from more familiar to less, always getting harder & requiring ever more brainpower, so you feel depleted much sooner and learn less. Doing it as above, you always start with the hardest bit, and it gets ever more familiar from there, so the exercise feels less depleting.

When you're memorizing a piece of music, the insider tip is *never* start learning from the beginning, always work from the end, for the reason above. And in performance, recording, performance-practice you'll playing it from the front on as per norm, so the backchaining and normal usage balance each other out. I'm not sure how frequently you're going to "perform" by listening to any one clip all the way though, though

I remember reading the theory of this somewhere (I think it was in the context of language), and I agree it makes a lot of logical sense. Although it will definitely have to be configurable. One thing that music and audio books have in common is that they tell a story and the order is important to help the meaning sink in in a comprehensible way. So despite the logic of back chaining, hearing things in their proper, meaningful order can help to aid memory as well. So for me a prerequisite to back chaining would be another training mode that gradually introduces you to the story in its intended order. Once you have that general familiarity with the story and how it ties together sequentially, you can go back and use a different training mode for memorisation (such as back chaining).

[Cainntear]

The back chaining I have used, both in Japanese and in music, goes:

D
CD
BCD
ABCD

That way, the material gets more familiar the further you go in. The problem with backchaining from the front is

... that it's not backchaining, because backchaining goes from the back.

that you're always moving from more familiar to less, always getting harder & requiring ever more brainpower, so you feel depleted much sooner and learn less. Doing it as above, you always start with the hardest bit, and it gets ever more familiar from there, so the exercise feels less depleting.

When you're memorizing a piece of music, the insider tip is *never* start learning from the beginning, always work from the end, for the reason above. And in performance, recording, performance-practice you'll playing it from the front on as per norm, so the backchaining and normal usage balance each other out. I'm not sure how frequently you're going to "perform" by listening to any one clip all the way though, though

I remember reading the theory of this somewhere (I think it was in the context of language), and I agree it makes a lot of logical sense. Although it will definitely have to be configurable. One thing that music and audio books have in common is that they tell a story and the order is important to help the meaning sink in in a comprehensible way. So despite the logic of back chaining, hearing things in their proper, meaningful order can help to aid memory as well. So for me a prerequisite to back chaining would be another training mode that gradually introduces you to the story in its intended order. Once you have that general familiarity with the story and how it ties together sequentially, you can go back and use a different training mode for memorisation (such as back chaining).

My issue with backchaining has always been that it is a rote memorisation technique -- the reason that you can't remember the whole word or phrase is that it's too complicated and contains too many new variables that you haven't acquired mastery of yet.

To me, the apparent success of backchaining in Pimsleur and similarly phrase-heavy courses is just proof that phrase-based learning is not appropriate for beginners.

So yes, I agree that configurable revision/repetition schemes would be a very good thing.

Just as long as the term "backchaining" only means "backchaining", cos otherwise it confuses people.

[tangleweeds]

Just wanted to self-indulgently add, for version 2.1 perhaps, that there may be nothing I appreciate more than being able to record myself trying to imitate a native speaker. It really helps so much with unfamiliar phonemes. I used it a lot in Anki to practice broad/narrow consonants in Irish, and now Lingodeer offers it too, which I'm using to practice pitch accent in Japanese.

[tangleweeds]

The back chaining I have used, both in Japanese and in music, goes:

D
CD
BCD
ABCD

That way, the material gets more familiar the further you go in. The problem with backchaining from the front is

... that it's not backchaining, because backchaining goes from the back.
<snip>
Just as long as the term "backchaining" only means "backchaining", cos otherwise it confuses people.

I must not have expressed myself well. I meant exactly this:
https://www.teachingenglish.org.uk/article/backchaining

[edit: Oops, I didn't read this right. I get you now. I was working with major sleep dep yesterday, we had had a veterinary emergency (cat repairs successful).]

My issue with backchaining has always been that it is a rote memorisation technique -- the reason that you can't remember the whole word or phrase is that it's too complicated and contains too many new variables that you haven't acquired mastery of yet.

I don't use it to memorize, it's more like learning tongue-twisters, learning to hear and reproduce unfamiliar phonemes and cadences. It's about listening and speaking skills, not memorizing content. Procedural not declarative.

The thing about interacting with real-live native speakers is that they invariably throw at you too many new variables that you haven't mastered yet. So that's actually a skill to practice under presssure.

[Apologies for the edits, sooner or later I always hit Submit instead of Preview.]

[ryanheise]

The back chaining I have used, both in Japanese and in music, goes:

D
CD
BCD
ABCD

I am currently trying to work out how this scales. My other two training routines scale well, but with back-chaining, does this chain just keep growing and growing, or do you break the chain after a certain length and start a new chain? And if so, to you then do some practice joining chains together? Since there will be options for everything, I'm just trying to understand how this idea could scale if people wanted it to.

What I may end up doing is doing what I had in mind first so you can get a feel for how it scales, and that might give you some ideas on how a similar thing could be done for back chaining.

[ryanheise]

Just wanted to self-indulgently add, for version 2.1 perhaps, that there may be nothing I appreciate more than being able to record myself trying to imitate a native speaker. It really helps so much with unfamiliar phonemes. I used it a lot in Anki to practice broad/narrow consonants in Irish, and now Lingodeer offers it too, which I'm using to practice pitch accent in Japanese.

One of the ideas I put up earlier in the thread was to record and then compare your voice to the original audio by analysing features of the audio that are important in speech (pitch and IPA sounds). It's definitely not on the short term plan but I think it would be pretty cool to provide this sort of feedback.

[jeff_lindqvist]

The music-style routine is this:

A,B,AB
C,D,CD
AB,CD,ABCD
E,F,EF
G,H,GH
EF,GH,EFGH
ABCD,EFGH,ABCDEFGH

And the routine I used for extending my listening attention span was something similar to this:

A,B,C,D,E,F,G,H,
AB,CD,EF,GH,
ABCD,EFGH,
ABCDEFGH

Somehow this got me thinking of the old HTLAL thread Strategy: Learn 600 words a week:

Divide these 100 flash cards, in 10 groups of 10 (at random). /.../
Memorization Order:

DECK (Number of Cards in this Deck)

A (10 cards)
B (10 cards)
AB (20 cards) <-- Which is A+B etc.
C (10 cards)
D (10 cards)
CD (20 cards)
E (10 cards)
F (10 cards)
EF (20 cards)
G (10 cards)
H (10 cards)
GH (20 cards)
I (10 cards)
J (10 cards)
IJ (20 cards)
AB+CD = K (40 cards)
EF+GH = L (40 cards)
IJ+K = M (60 cards) <-- Which is IJ + AB + CD
M+L = N (100 cards) <-- Which is IJ + AB + CD + EF + GH

But even if I responded to that thread (back in 2007), I can't remember if I ever tried that method.

Fast forward to 2020 and your July topic, I wonder if we memorize (and recall?) things differently, and/or if it depends on the type of content (music, dialogue etc.)

Although I can't reproduce, say, an Assimil dialog today, I could do it so-so some years ago when I was shadowing the content. But I don't really call that memorizing. It felt more like I was learning to retell a story, and that is also how I learn music. (I still don't say that I memorize it - that's something else.) Each phrase made sense and lead to the next one. Like music.

Olle "Speech doctor" Kjellin says:

(...)just consider how much easier it will be to learn a new song if you are allowed to sing along a number of times with someone who knows it, compared to having to try solo singing after only having listened to it a few times.

(Source: http://olle-kjellin.com/SpeechDoctor/ProcLP98.html )

In your Experiment 5, you say that you mimic what you're hearing. That's shadowing to me, and I actually think that's the only thing I'd need to "memorize it" (gaah, I don't like that expression). As long as you understand what they're saying (or, as I "understand" what they're playing), just repeat aloud. I'd even go so far and say that those who don't understand the music (or the language) are those who need (or think they need?) to memorize it bit by bit. Play the melody, not the notes.

These are my own thoughts - totally unscientifical.