suegy / abode-star Goto Github PK

The validate primitives button should (in theory) load a behaviour primitive 
file from a Python or Lisp file. Checking against my Python behaviour 
libraries, this doesn't seem to work.

Currently it always says all of primitives are invalid.

What is the expected output? What do you see instead?
Actually get the primitives from the behaviour file.



Not actually convinced this is a hugely useful option even if it did work, 
probably not worth making this usable?

Currently there is no undo or redo functionality in the current version of 
ABODE.

Desirable if it was possible to undo actions such as:

Creating new nodes.
Deleting nodes.
Changing node properties.
etc.

What steps will reproduce the problem?
1. Set a competence to be disabled
2. Save a file
3. Reload the file

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

When loading the file again, the parser should in theory (from what I have seen 
in the code) should load the commented out sections (disabled competences is 
just commented out) however this doesn't work and is just ignored.

So this is not loaded correctly.

The print button doesn't do anything, literally has no code in the pressed 
event. Disabled this for now.

What steps will reproduce the problem?
1. create a new posh element eg. an action
2. use undo 
3. switched tab and changes sometimes parts of the lap

What is the expected output? What do you see instead?
delete the element, stay in same tab

Currently ABODE will overwrite any manually added comments to a POSH action 
plan, which means that it is hard to develop a robust plan that is human 
readable outside of ABODE.

Ideally changes should be able to be made to a plan within ABODE, without 
overwriting comments.

In the current version, there are multiple input boxes which are text input, 
which should be replaced with drop-down menu or any other viable alternative.

These are text when there is only a very limited amount of valid options, which 
means a drop-down menu / radio buttons would be desirable.

Comprehensive list of offending input boxes

In Drive-Element properties:
  Frequency Unit (seconds,minutes,hours)

In Competence properties:
  Timeout Unit  (seconds,minutes,hours)
  Enabled (true, false)

In Sense properties:
  Predicate ( = == != < <= > >= or blank / default)

What steps will reproduce the problem?
1. Build Drive and DriveElement
2. Try to add an action
3. The only choices currently offered are ActionPattern or Competence

What is the expected output? What do you see instead?
It should be possible to add any of the three elements in an interchangeable 
way: Action, ActionPattern, Competence

The size and position of the command buttons (top right buttons, used for 
moving / deleting / merging nodes) change depending on the currently selected 
nodes. This is a major design flaw, as a user can easily press a "Move node up" 
button multiple times, only to find that the button has turned into "Delete 
node" after the first click. 

Without undo functionality, this is a noticeable usability issue. 

When closing ABODE, there is currently no prompt when you have an unsaved file 
open. The program will just close, potentially losing large amounts of unsaved 
work.

Ideally, when closing there should be a prompt for:
"Save" "Don't Save" "Cancel"

When clicking on an element that does not have actions associated with it (such 
as "Drive Collection" and "Competence") the last set of available actions 
remain on screen, but do not do anything.

Instead, these buttons should ideally disappear / be grayed out.

What steps will reproduce the problem?
1. open plan file
2. delete elements and add elements
3. save plan file


What is the expected output? What do you see instead?
additional top level brace appears in file which is invalid

Using the editor should allow for at least a single lap file or preferably a 
whole agent using multiple lap files the visualization of used actions and 
triggers to visualize the action selection process.

What is the expected output? What do you see instead?
Highlighting of regions and elements regarding their activity during run-time 
of as a heat map off-line.

Buttons and frames are mostly to small to use frames sometimes even do not show 
their content until you click on them or try resizing them.
After resizing the correct size is used.

What is the expected output? What do you see instead?
all elements should be shown in full size. If frames do not fit we should  
think about minimizing certain tabs to remain optimal size of others.

As far as I am aware, there isn't much in the way of documentation on how to 
use ABODE.

The User Manual section in ABODE is currently empty.

It might be worth putting a quick-start guide on how to make a basic LAP file, 
as well as briefly describe the different elements of POSH. 

What steps will reproduce the problem?
- delete button for action pattern and competence behave differently 
- each posh element is re-implenting a lot of the same functionality again and 
again

What is the expected output? What do you see instead?
same behaviour for buttons with the same label and same expected functionality

Please use labels and text to provide additional information.

What steps will reproduce the problem?
1. create an AP and add new DE
2. add Competence "a"
3. set trigger for "a" to "a"

What is the expected output? What do you see instead?
It should not be possible to create circles because these mean infinite 
sublists (which is bad!!!)
Before setting a trigger it needs to be checked that this does not create a 
circle (a bit tricky)

Would be ideal if there was some drag and drop functionality. The current 
system is not particularly intuitive, so this would be a bit more usable 
technique.

e.g. being able to drag nodes onto others to set that node as a child. 

As part of this, it would make the use of the program more obvious to a new 
user if they could see what node types were available as part of the tool bar 
and then drag them into the workspace. 

The tree display at some point in the past was changed so that it displays the 
actions before the triggers that cause the action.

See attached example screenshot.

This should probably be changed back to the old way, as it made more sense. 

Ideally would like the ability to rename all instances of a sense / action 
element name.