A new approach for Variable Naming for the workflows is needed

[lucasaugustomcc]

In my opinion, the current approach for variable naming is a little confusing.

For each input variable of the notebook, an informative name will be generated for the workflow as follows.
The name will start with:

• “I” if it is provided by the user and is not a variable with a primitive type (since those become parameters as explained above);
• “P” if the variable is a parameter;
• “D” if it is already a file in the notebook;
• “V” if the variable is passed from previous cells;
• "O": lacking mention and explanation about it in the paper (see figure below).

The name will then include the ordering number, and finally the name of the variable in the notebook.

For example, the third user-provided input for the fifth notebook cell would be named I.00003.c5.

My suggestion is to use the name of the variable in the notebook and only add a number to this name if there is another variable with the same name (i.e., data will become data2, since there is already a variable named data in the workflow).
Since we are assuming that NiW will create the workflow template automatically, I think the use of "I", "P", "V", "O" and etc, starting the variable name is no longer needed.

Is there any scenario/case that would require the ordering number added to the variable name?

[reginawang3495]

I think this would work out fine though you may want to separate D and I since for D, the user will need to input the file themselves (like "d.txt" in DiseaseAnalysis). For I, it would be like an input array so that the user can change the array without having to change the code. Of course, since the I variable was inside the code in the first place, the system will automatically generate a file for the original input (in this case the original array). However, there is no case that would require the ordering of the number.
I implemented the IPVOD originally because I thought that Wings took in variables by alphabetizing them before putting them in the "sys.argv" array(hence the O00001graph instead of O1graph), but since it is unnecessary, you can remove it.

[reginawang3495]

I just noticed this, but its not just I0001c1 or completely depending on the position. It uses the name set by the programmer. So if you see in FullDiseaseAnalysis in wings workflows, all the names in there were provided by the system
So if there is a string provided by the user and it could be P0001string or P0002blah or whatever depending on what the user named it AND the position (so number is by position, P is for parameter, and the end text is the name the programmer gave)