What steps will reproduce the problem?
1. Make an exploration with lots of nodes
2. Try to look at the exploration graph

What is the expected output? What do you see instead?
I would expect for the graph to be scrollable or resize itself when I run out 
of space. instead it just kind of goes out of bounds.

I wanted to make a multiple choice answer that had the Spanish word marrón in 
it, but it wasn't allowed.

Once an exploration gets pretty big, it's a bit cumbersome to find the needed 
state in the dropdown list when you are defining a new rule.
It would be really cool if instead of a linear dropdown, we could have a mini 
version of the exploration graph pop up, and I could just click on a state 
there.
Then I guess we would also need a separate "new state" button. Or we could 
render the exploration graph with a new state already added, connected to the 
state I'm making the rule for, and if the user clicks on that one, *then* 
create it for real?

What steps will reproduce the problem?
1. Create a new exploration.
2. Enter the State Editor for any of its states.
3. Add a text item to the non-interactive content section and click 'Edit'.

What is the expected output? What do you see instead?
Only the toolbars are shown. There is no space to input the content.

What version of the product are you using? On what operating system?
Firefox; Linux Ubuntu

What steps will reproduce the problem?
1. This one can be a bit flaky but I found that sometimes on a line my 
backspace key moves my cursor to the start of the line

What is the expected output? What do you see instead?
Cursor should only remove the previous character. Instead the cursor moves to 
the start of the line

What version of the product are you using? On what operating system?
Mac OS X Mountain Lion (10.8.2)
Chrome  25.0.1364.172
I checked out the most recent source and ran my server locally.

Please provide any additional information below.
I submitted another issue as [email protected] I will keep an eye on that 
email but generally I will be using this one. That other email was a mistake :-)

While Installing oppia I have the following error
scripts/test.sh: line 103: third_party/node-0.10.1/bin/npm: Permission denied
I followed the exact same procedure but instead of getting my localhost started 
I am getting the above error. 


What version of the product are you using? On what operating system?
I am using npm version 1.4.1 and downloaded the latest oppia code from git. 
Ubuntu 12.04 LTS

It would be cool if the content creator could specify certain assertions that 
communicate intent of flow, and as part of the sanity checks we checked whether 
these assertions were true.
For example:
There can be lots of cases in the exploration where if you have arrived at a 
certain state, we assume that you've already been to another previous state or 
have already learned something. For example, the state might say "as we already 
learned..." or "but we already know that..."
It would be nice to have a way of asserting that, so that in an exploration 
with a complex flow, this dependency always stays true and doesn't get messed 
up as the flow gets more tangled.
(we would probably want to make it a warning instead of an error)

This is something my mom noticed. The "Type your answer here" text stays in the 
text input box even when you click into it, so she wasn't sure if she needed to 
delete it somehow before she started typing, and it was somewhat unnerving to 
her.

Our parameter model is limited and a bit unclear, and this needs fixing. Please 
have a look at the proposal at:

    https://code.google.com/p/oppia/wiki/ParametersDesignDoc

and leave any comments you have there (especially Yana and Koji, since you have 
been running into issues with parameters recently!).

Once we agree on the correct plan of action, I will use this issue to track 
progress on implementing this proposal.

Render Interactive widget w/o interactions for reader.  Used to show the answer 
user provided. Two use cases:
- Final (correct) answer in END state. E.g. map widget with a marker for the 
correct answer (as well as all the incorrect answers user provided). It can be 
interactive in the sense view port and zoom can be manipulated, but clicking on 
it should not send the answer any more.
- Show the answer user provided when the new state uses a different widget (or 
the same widget but sticky config is false). This requires adding a new iframe.

In addition to leaving the widget rendered (and maybe create a new iframe for 
the new interactive widget), we need to send a method to the widget so it can 
update it's rendering and stop sending answers.

What steps will reproduce the problem?
1. Complete a lesson

What is the expected output? What do you see instead?
I expect to see: Congratulations, you've finished this exploration!
I actually see: Congratulations, you\'ve finished this exploration!

It looks like you are escaping the ' which does not seem necessary

What version of the product are you using? On what operating system?
checked out lastest source
Mac OS X 10.8.2 (mountain lion)
Chrome 25.0.1364.172

Please provide any additional information below.
I actually just made a patch for this. Hopefully at some point I can make a 
more significant contribution ;-)

I think it should be easier to get from a state to "Neighbor" states, i.e. 
(1)states that this state can lead to, and (2)states that can lead to this 
state.
For (1), it could simply mean making the "destination state" listing in the 
rules be a link to the state in question.
For (2), maybe just a list of states up near the top? maybe even with a brief 
indication of the rule that leads from the state in question to this one?

What steps will reproduce the problem?
1. create new exploration
2. click on the first state created by default, and rename it
3. Go back to the exploration graph

What is the expected output? What do you see instead?
I would expect the new name to show up in the graph. Instead it still says 
"[untitiled state]", probably because it's being cached client-side somewhere 
and that particular cached version does not realize it's been renamed

Please use labels and text to provide additional information.

Unclear which widgets are impacted or exactly where the issue lies, but I've 
seen several different behaviors including records for answers never input and 
activities I've completed that have no recorded answers.

It would be very useful to have rules that are Boolean combinations of other 
rules. In other words, it should be possible to construct rules using other 
rules together with AND, OR and NOT (and perhaps ANY).

This could be important for language explorations.

For example, in English,

"he and i went to the store." may be classified as wrong while
"He and I went to the store." might be classified as right.

From Sean's mail:

> 5. I now know at least five different people who are interested (in different 
contexts) in the idea of having a student enter stuff into a programming 
console and get interactive feedback on his/her program. Would any of you be 
interested in looking into how Oppia could more directly support this use case? 
Basically instead of trying to interpret strings and numbers we would now have 
to try and deduce things about programs (a simple example would be which test 
cases they pass).

What steps will reproduce the problem?
1. launch (local) oppia server
2. In the state editor, click "add Image"
3. Select image and click "save"

What is the expected output? What do you see instead?
Nothing happens. I would expect the image to get "uploaded" somehow or at least 
an error message to show.

Currently all rules classify based on the reader's answer only.

However, it would also be useful to have these rules classify based on 
parameters associated with the reader. This allows classification on things 
such as:

(a) how many guesses has the reader made on this particular question?
(b) has the reader learned, demonstrated knowledge of, or demonstrated 
proficiency in, skill X?

The graph and tree visualizations have a fixed height. However, the height of 
the svg ought to change as the number of nodes increases.

Actually, it probably shouldn't matter whether the exploration is published or 
unpublished, but when the content creator clicks the play button to test out 
his own exploration, those tests should probably not be logged as actual "uses" 
of the exploration, and shouldn't log stats.

What steps will reproduce the problem?
1. Create some text in the rich text editor that spans multiple lines
2. Highlight the lines
3. Change the font

What is the expected output? What do you see instead?
I would expect the entire block to change font. Instead I have to highlight one 
line at a time and change the font.

What version of the product are you using? On what operating system?
Mac OS X Mountain Lion (10.8.2)
Chrome  25.0.1364.172
I checked out the most recent source and ran my server locally.

Please provide any additional information below.

It would be cool if the list of parameters in the Exploration Metadata also 
showed where (in which states) each parameter is changed and/or used.
Not sure how to best do this though.

What steps will reproduce the problem?
1. Create a parameter change
2. Type in a non-trivial jinja expression, like {{Num}}+1
3. Click save
4. Try to edit the parameter change

What is the expected output? What do you see instead?

The chip displaying the existing value should read the same as what you typed 
in, e.g. "{{Num}}+1"

instead, it tries to be smart and extract the name of the variable, so it only 
displays "Num"

I think this does not actually affect the value, just what the user sees (but 
not sure)

Especially if the user has no explorations yet. I think it is a bit 
counter-intuitive, now that we have that page, to only have the "create new 
exploration" button in the Gallery.

Allow multiple classifications to go to the same feedback set.

For example, you might want the answer to be one of [Adiós, Adiós!, Ciao!, 
Ciao] but you don't care which one they use.

This can be worked around for now by adding multiple classifiers with the same 
info, so it isn't high priority.

Currently, parameter changes can only be specified at the start of a state, and 
at the level of an exploration. There should also be a UI that allows parameter 
changes to be specified based on the rules triggered by a reader.

What steps will reproduce the problem?
1.Try to create a new state.
2. Get this message in the butter bar:

Warning! Server error when adding state: convert_state_name_to_id() takes 
exactly 2 arguments (1 given).Please refresh your page.

3. Refresh page; new state is actually there

Right now, looping back to the same state is treated differently than state 
transitions to different states: the state text does not get re-displayed, and 
the param changes do not get re-evaluated. This makes sense in most 
applications so far, since most of the time a looping back represents a lack of 
progress, but sometimes the looping back is actually meant to represent a 
"true" transition from a state to itself.
For example:
suppose I'm making a binary search game, where the exploration tries to guess 
what number you are thinking of. The central state in this exploration* updates 
the min and max of the range of possibilities, then picks the number in the 
middle of that range, and guesses it. The player then says if his number is 
smaller or bigger than the guess, and the state loops back. In this loop back, 
I do indeed want to update the max, min, and guess, and re-display the text 
(i.e. make another guess)

I think there should be a check box that allows the loop-back to act as a 
"true" state transition.


*this is actually a bit of a simplification: I had to copy this state into 
three, because I can't do parameter changes in state tranistions, so I had to 
have additional "last guess is smaller" and "last guess is bigger" states.
But I still have to loop back in roughly half the cases.

What steps will reproduce the problem?
1. Define a new parameter change in a state
2. go to the exploration, and click on "show exploration metadata

What is the expected output? What do you see instead?

I would expect the new parameter to show up in the list of parameters.
It does not show up until I go out of the editor and come back (and the 
exploration is reloaded, I assume)
I also can't add the same parameter to the exploration-level list manually, 
haven't investigated why yet

Looking at answers we record, I came across a record for a MC question that had 
just the response '3' which was not helpful. Instead we should be recording the 
text.

Render Interactive widget for preview in editor.
- Do not send answers.
- Keep the configuration -- e.g. view port and zoom for the map widget.
- Default to the normal interactive rendering.

Need a exploration that has non-sticky map widget to repro this.
This should be a problem for any iframe'ed static rendering of widgets (but 
InteractiveMap is the only such widget now).

The parameter changes should be reorderable, just like the classification 
rules, in case the order matters.

SImilar to issue #29, but instead of just making sure there's a path from start 
to end, make sure there's a path from Node X to end, for any Node X. 
Something like: traverse the graph from END backwards along the transitions, 
marking each node that's reachable as "OK", and then mark any nodes not marked 
as "OK" as broken?

It breaks on the "sed...Cakefile" commands.
I also briefly tried "sudo apt-get install cakephp-scripts" as recommended in 
the comment, but of course apt-get is not a thing that works on mac either.
Didn't immediately find a way to get the equivalent scripts on mac.
The good news is that everything else works! (after disabling jsrepl and 
manually installing mac version of node.js)

I made a rule a while ago to classify on multiple choice input. It has been 
working just fine. I edited one of the rules to point to a different state. 
This still seems to have worked correctly, except that in the editor, the rules 
themselves are showing "undefined" and the dropdown is not populated with the 
multiple-choice choices when I try to edit the rules. (screenshot attached)

note: I did accidentally overwrite the comment-out edit that disabled whatever 
wasn't letting me add more states. Could this be a related bug?

ERROR    2013-03-15 18:28:48,066 webapp2.py:1552] argument of type 'int' is not 
iterable
Traceback (most recent call last):
  File "/home/sfederwisch/oppia/oppia_runtime/google_appengine/lib/webapp2/webapp2.py", line 1535, in __call__
    rv = self.handle_exception(request, response, e)
  File "/home/sfederwisch/oppia/oppia_runtime/google_appengine/lib/webapp2/webapp2.py", line 1529, in __call__
    rv = self.router.dispatch(request, response)
  File "/home/sfederwisch/oppia/oppia_runtime/google_appengine/lib/webapp2/webapp2.py", line 1278, in default_dispatcher
    return route.handler_adapter(request, response)
  File "/home/sfederwisch/oppia/oppia_runtime/google_appengine/lib/webapp2/webapp2.py", line 1102, in __call__
    return handler.dispatch()
  File "/home/sfederwisch/oppia/oppia_runtime/google_appengine/lib/webapp2/webapp2.py", line 572, in dispatch
    return self.handle_exception(e, self.app.debug)
  File "/home/sfederwisch/oppia/oppia/controllers/base.py", line 144, in handle_exception
    webapp2.RequestHandler.handle_exception(self, exception, debug_mode)
  File "/home/sfederwisch/oppia/oppia_runtime/google_appengine/lib/webapp2/webapp2.py", line 570, in dispatch
    return method(*args, **kwargs)
  File "/home/sfederwisch/oppia/oppia/controllers/base.py", line 81, in test_editor
    return handler(self, user, exploration, state, **kwargs)
  File "/home/sfederwisch/oppia/oppia/controllers/editor.py", line 355, in put
    if ('{{' not in rule['inputs'][param] or
TypeError: argument of type 'int' is not iterable

For the interactive map, classifiers need to be specified by latitude / 
longitude coordinates and distance all in numbers.
It'd be nice to allow adding classifiers with interactive widgets, e.g. draw a 
circle in a map widget to specify the correct answer region.
It'd be even nicer if we can specify multiple circles on a same map widget for 
specifying odd-shaped region (or allow specifying polygons, but that'll make 
the readers side logic more complex). 

Like if nothing points to the end state at all. Or if it is not in the set of 
states reachable from the start state. I think we already mark such unreachable 
states as red, so it probably wouldn't be *too* hard?

What steps will reproduce the problem?

1. Open a multiple-choice widget customization editor.
2. Edit an item.
3. Click on the 'Save' button for the modal without saving the individual item.


What is the expected output? What do you see instead?

The modal 'Save' button should be grayed out -- I should need to click on the 
item's save button before I can save the customization parameters as a whole.

from review comments:

https://code.google.com/p/oppia/source/detail?r=ba98251953359e2a94fdb7bc20a4cdf1
5a16ab49

instead of one reader_response.html that works for all widgets, make a copy in 
each widget directory and customer there.

Roughly:
- Make a google code account (or assume it's there).
- Clone oppia.
- Copy to local.
- (make changes).
- Commit and push to the clone.
- Make a pull request.
- Review cycle?
- [Oppia maintainer] Pull the change to main repository.

What steps will reproduce the problem?
1. Create a new parameter change
2. For the values, try to type in a jinja expression like "{{[Num, 1]|sum}}"

What is the expected output? What do you see instead?

I would expect that to be parsed as one expression. Instead, it is parsed as 
two potential values:  "{{[Num" and "1]|sum}}"


Please use labels and text to provide additional information.

Another generic interactive widget which should be useful in many contents.

Need to be able to specify an image (perhaps a image url).
Classifiers would be click regions -- similar to InteractiveMap's, but the 
coords should be image size normalized to [0, 1] range.
Consider allowing it to be sticky, but this needs rendering the answered region 
somehow on the image.

Sometimes it makes more sense in the flow of the interaction for the player to 
make a choice by clicking on a button instead of selecting a choice, then 
clicking submit. Or sometimes I just want to tell force the user into an 
"action", without giving him an interaction choice.
Examples:
"OK, now think of a number", and then have just one button that says something 
like "I thought of a number" or "Got it" instead of "Continue".
Or, "Would you like to think of a number for me to guess, or try to guess my 
number" and have two buttons, one for each choice.

What steps will reproduce the problem?
1. git pull oppia on a mac
2. run start.sh

What is the expected output? What do you see instead?
The start.sh script pulls the linux version of node.js and npm without looking 
at the actual system, so when it tries to actually use it to get the needed 
node packages, it breaks since the binary can't run on mac.

The content creator can specify several options for the feedback that the 
reader receives. If the reader hits that rule, Oppia then chooses one of the 
feedback options at random.
In the case when the content creator specifies more than one option, these 
options are often hints of some sort, and allow the reader to "try again" with 
more information about the problem. In this case, it might make more sense to 
choose these options without replacement, i.e. once an option has been shown 
once (in some context/scope), it is not shown again until all the other options 
are exhausted.

I think the problem with this suggestion is defining the context/scope in which 
the "without replacement" part would occur, and outside of which the 
memorization of which options have already been shown is reset. So, for 
example, if the player moves away from that node for a while and then somehow 
comes back to it during the same session, but much later, Oppia should probably 
"reset" the memory of which hints have already been shown.

I think a nice way to do this would be to say that Oppia "remembers" which ones 
have already been shown until the reader breaks the "streak" of hitting that 
rule and hits a different rule. Although that might be a bit too sensitive, 
since there could be several "no that's wrong" rules that give similar hints, 
so alternating hitting those rules might erroneously reset the "memory"
Or maybe it should reset when the reader moves away from the state.

What steps will reproduce the problem?
1.Start editing one rich text box
2.Forget to click save and get distracted
3.Start editing a second rich text box in the same state

What is the expected output? What do you see instead?
Both are open for editing, but in weird places, and neither saves properly. I 
would expect either:
a) pop-up error that won't let me open the second one
b) first one gets auto-saved before second one opens
c) everything works magically and I can edit both at the same time

When the content creator is testing out his own exploration, suppose he 
discovers a flaw in the middle of it. it would be nice if he could click on a 
"edit this state" button and go directly to the state he is on.