Reading and Word Meanings For Beginners (Design Discussion)

Demo for only one poem:

https://math-eight.blogspot.com/2025/03/twinkle-twinkle-little-star-word-by.html

On three clicks of "Next":

On one click of "Previous":

When it switches from word to word: it speaks that word out (No magic, just Text to speech in browser).

Ques: WHERE WE ARE COMING FROM?

This was the first app of translation:

We received feedback on this from various students from various backgrounds.

The situation was like this:

There could be three types of users/students of this app:

1. Those are not able to do even single sentence (due to no capacity for reading even simple English words). These students would want something easier, or lighter on them, such as a "Reading App" - like the one we are designing right now.

2. Those who are able to take on these translation questions albeit with some difficulty -- Those are actually learning from this app.

3. Those who already excel in this type of translation questions and who may want to pursue something tougher like the grammar, story creation.

Limitations

1. 

Subhashish: Support for multiple languages where it is showing pronunciation and meaning

2.

Akash:

- There is only one poem. 

- Limited dataset / library

3. Akash:

This not targeting any age group (5th or 6th or 8th) or learner level (beginner, medium difficulty for someone trying to build his / her vocab).

------------------------------------------------------------------------------------------------------------------------

My comments:

1. We are doing a POC (Iteration 2) so we have limit our scope.

Scope could be support for a third language (of native speakers) such as Bengali.

How are we putting together these word meanings:

We are pulling them out from Gemini (LLM by Google).

So if Gemini also understands Bengali then yes, we can do it.

Subhashish would have to pitch in for basic sanity and testing of Bengali sections.

2. Yes, there is only one poem at the moment but that was because we were doing a small POC.

And secondly because: Who decides what poems to put in here? You, me or the users, or the teachers, who?

I thought that let's have users or teachers decide what chapters to put in here.

And that would be made dynamic in the mobile using following methodology:

2.1. User would click a picture of the page from his or her book.

2.2. An image to text AI model would pull out text from that captured image.

2.3. Then a GenAI model such as Gemini would be used to collate the word meanings for the text from the image.

2.4. This metadata about the poem/chapter/ or text in general would be saved on the device of the users for offline reading.

Here Steps 1 to 3 would require internet.

3: Yes, Akash is right. We are not targeting any age group. We are rather targeting users who do not even read or understand basic English words. It could be someone who is in 3rd grade and trying to improve his/her English. Or it could also be an uneducated adult trying to gain some language skills today.

==========================================================================

Couple of features I am thinking of keeping in it are:

1. This would fully support online and offline engagement.

If student wishes to download all the chapters from the database onto the device itself and then use the app offline, this option would also be available.

2. A feature that would be supported by only online engaged users will be ASR: Automatic Speech Recognition for live reading and correction.

3. Free and open for all.

4. Database: SQLite in mobile for offline users and MySQL via PythonAnywhere for online users.

==========================================================================

Homework given on 2025-Jul-18:

Subhashish: 

1. How would you design this app's front end?

2. How would you design the backend? 

Akash:

1. How would you design the processes and programs that would enable the backend for this app?