possibly due to efforts to update/migrate to v2

Improve it more by following research-based best practices: https://designguidelines.withgoogle.com/conversation/conversation-design/welcome.html

fix bug with stray variables or variables not getting recognized

Don't ask what would you like to create, change all to ask which would you like to create?

And maybe change all intros accordingly? (Variations are still possible, and would be good for keeping things fresh for repeat users.)

repeat 3 times
say hi

should output

hi hi hi

see #9

a = 1
say a
a = 1 and 2 and b

should output

1
[1, 2, "b"]

see #9

should still work even with #21
test:

if 1 equals 1
say hi

should output

hi

see #9

(needs to be an extra intent - not because of different recognition pattern but because of an extra value extracted)

Consistency. Also let the user know the code ended.

a = 2
if 1 equals a
say hi

should output

see #9

if 1 equals 2
say hi

should output

see #9

Awaiting the Google Assistant app directory to update. The app's already approved. Talked to AoG support already and they're still working on it. For now, the quick links in this repo are down until they get finish reviewing.

Action item once they're done: double-check both quick links in this repo work.

Apparently the mic is sometimes still on with no prompt when the user says "nothing" or "no".

From feedback: (not sure if this is the good/bad example)

• Action: What's the secret word?
• User: Nothing
• Action: Code Tutor version 4, signing out. Come again for updates.
• Code Tutor left the conversation

Or:

• Action: What's the secret word?
• User: No
• Action: Code Tutor version 4, signing out. Come again for updates.
• Code Tutor left the conversation

Things to check:

• does the secret word request actually respond with nothing and still is listening to "nothing"/"no"? (might be timing bug, not my code's problem)
• otherwise, if it responds with "Code Tutor version 4, signing out. Come again for updates.", then that is intended. but if the mic is still on, then check if it properly exits. might need to set custom js fulfillment code to exit.
• are there other intents that are listening for "nothing"/"no"?
• if it does properly exit (i.e. with message and with mic off), then get more info/feedback

whoops

Make use of those multi-line responses of "images from text".
Cards/suggestions can be like buttons that are activated by voice.

Tic-tac-toe?

a = 1 and 2 and b
say a

should output

[1, 2, "b"]

see #9

(Based on some of the feedback I've gotten so far.)

Idea 1:

Give a code example to set expectations?

Right when a user "opens" the app, set expectations by showing an example of code?
(So change the intro?)

For Example:

Hi! This is Code Tutor.

Here's an example of what you can make:

let greeting = "hi there ";
let name = "someone";
let message = greeting + name;
if (name == "someone")
  say("What's your name?");
for (let i=0; i<3; i++)
  say(message);

Voiceover:

Hi! This is Code Tutor. Here's an example of what you can make:
Let greeting equal hi there.
Let name equal someone.
Let message equal greeting plus name.
If name equals someone, then say what's your name.
And then, while counting from 0, and stopping before 3, say message.

a = 3
repeat a times
say hi

should output

hi hi hi

see #9

(ascii art is broken)

need to implement "run code"

Some cases to consider (not exhaustive):
ignore code without say
ignore code after last say

a = 1
say a
a = 2
say a

should output

1 2

see #9

"and" is becoming "+" with voice recognition

i.e.

if/for ...
if/for ...

becomes

if/for ...

if/for ...

a = 1
say a

should output

1

see #9

Here's what you can say: (read the list out loud)
(show list of things, including run code)
If you need this list again, just say "what's on the list?"
V
(repeat fill and transpile)
V
(run code? interpreter runner!)

handle cases:
x equals 1 plus 1
x equals a plus a

Add note in exit message? "Come back for updates."

a = 1
if 1 equals a
say hi

should output

"hi"

see #9