 |
Author |
Topic: well (Read 1769 times) |
|
benelek
 
 
|
Re: well
« Reply #15 on: Mar 11th, 2003, 12:43am » |
|
back onto the topic of code visualisation... do u realise that the code colouring, and "beautify" tool in P5 is a start in this direction? perhaps a more involved editor could use expandable tree-structures to view functions, methods and objects... and in the future, who knows - maybe even a way of editing code in a 3D web of code instead of a tree structure? code structure that visually and interactively behaves like the underlying concepts of a language (ie, object orientation, multithreading, etc)?
in terms of this code visualisation, it seems important that ur not just displaying it in an interesting way to a viewer; ur actually keeping it between the designer and the program, and moving in the direction of intuitive coding.
-bnlk
|
|
|
|
|
mmuday
|
Re: well
« Reply #17 on: Mar 13th, 2003, 10:21pm » |
|
I agree that P5's beautify tool is a step in that direction. One additional step in that direction would be providing a sort of "visual hungarian notation", where each declared variable/member has a unique set of visual attributes. I would like to do more than just changing the color of the text... at some point, it just becomes text soup and the visuals stop being all that meaningful.
For example, a String variable would have a specialized icon associated with it (maybe showing up above/below/beside/behind the variable name), perhaps with some particular coloring that differentiated it from other string variables in that method.
Furthermore, it would be cool to visually differentiate members from local variables.
Does anyone else think this would be an interesting step Does anyone care to help me come up with some specifics on this one
What I'm envisoning as a first step is a Processing read-only pretty-printer which shows code with the enhancements mentioned above.
What I'd eventually like to end up with this particular experiment is a sort of visual narrative view of code. My hypothesis is that maybe by using visual hooks, code could be more understandable than just with plain-ol' text.
Ideas, anyone
-Mark
on Mar 11th, 2003, 12:43am, benelek wrote:back onto the topic of code visualisation... do u realise that the code colouring, and "beautify" tool in P5 is a start in this direction perhaps a more involved editor could use expandable tree-structures to view functions, methods and objects... and in the future, who knows - maybe even a way of editing code in a 3D web of code instead of a tree structure code structure that visually and interactively behaves like the underlying concepts of a language (ie, object orientation, multithreading, etc)
in terms of this code visualisation, it seems important that ur not just displaying it in an interesting way to a viewer; ur actually keeping it between the designer and the program, and moving in the direction of intuitive coding.
-bnlk |
|
|
|
|
|
|
mmuday
|
Re: well
« Reply #18 on: Mar 13th, 2003, 10:29pm » |
|
That looks really interesting... how did you stumble upon it
on Mar 11th, 2003, 4:24am, Martin wrote:
|
|
|
|
|
_martin
Guest
|
Re: well
« Reply #19 on: Mar 14th, 2003, 12:53am » |
|
on Mar 13th, 2003, 10:29pm, mmuday wrote:| That looks really interesting... how did you stumble upon it |
|
had a chat on visual programming with a fellow student here at the ateneo and did a little research. talked about 2D programming languages (programs do not have to go from top to bottom), among others. in short, googleit. 
|
|
|
|
|
benelek
|
Re: well
« Reply #20 on: Mar 14th, 2003, 7:07am » |
|
mm, i don't know about assigning icons to variable types, etc. how does that create any more meaning than simply colouring them differently?
i was thinking along the lines of displaying a warped 3D java heirarchy, as different parts of the language are used.
for instance, each time java.lang.Object is used to define a new Object, that part of the java heirarchy would be warped in. the inherited methods would fall into the new object's "bag of tricks", branching off bellow the new object.
the differentiation of local and global variables, etc, would then naturally be visible based on the variables' relation to the part of the heirarchy closest.
does this make sense?
-jacob
|
|
|
|
|
|
