Global or local variable (in the SIC sense)?

SicVar and SicStructure instances have a __siclevel__ attribute which is an integer set to the level of the corresponding variable in SIC (0 for global, 1 for first local level, and so on).

When SIC defines any variable, it automatically imports it in the Python __main__ name space¹². When this variable is SIC-local (when executing procedures for example), importing it in Python may overwrite a lower level instance which has the same name. To prevent this, the Sic object in Python __main__ provides an array of 10 instances (one per level) which are used to temporarily store object which would have been erased. This array is named localspaces and objects are saved as attributes of its 10 elements. Consider these SIC procedures:

SIC> type localtest1.sic
define integer a
let a 123
pause
@localtest2.sic
SIC> type localtest2.sic
define real a[3]
pause

They will define local variables named `A' at different local levels. Let's define a global `A' variable and execute these procedures:

SIC> define char a*8
SIC> let a "qwertyui"
SIC> @localtest1.sic
SIC_2> define integer a
SIC_2> let a 123
SIC_2> pause
SIC_3> python
Entering interactive session. Type 'Sic()' to go back to SIC.
>>> a
123
>>> a.__siclevel__
1
>>> Sic.localspaces[0].a
'qwertyui'
>>> Sic.localspaces[0].a.__siclevel__
0

`a' object currently in Python __main__ is the level-1 SIC object. The SIC-global `A' variable has been saved as an attribute of the Sic.localspaces[0] element.
Let's go deeper:

SIC_3> continue
SIC_2> @localtest2.sic
SIC_3> define real a[3]
SIC_3> pause
SIC_4> python
Entering interactive session. Type 'Sic()' to go back to SIC.
>>> a
[ 0.  0.  0.]
>>> a.__siclevel__
2
>>> vars(Sic.localspaces[0])
{'a': 'qwertyui'}
>>> vars(Sic.localspaces[1])
{'a': 123}

Now we find the level-2 object in the Python __main__, and the 2 lower-level variables with the same name are saved in the Sic.localspaces array. At any stage, user can access its current-level variables in the Python __main__, but also any lower level ones through the Sic.localspaces array.
Let's end procedures:

SIC_4> continue
SIC> python
Entering interactive session. Type 'Sic()' to go back to SIC.
>>> a
'qwertyui'
>>> vars(Sic.localspaces[0])
{}
>>> vars(Sic.localspaces[1])
{}

In an automatic cascading mechanism, the lower-level variables are moved back to the Python __main__ when the current level one is deleted. The Sic.localspaces array is cleaned at the same time so their is no double reference for any object. Thus at the end of the procedures, the SIC-global variables are back in the Python __main__, and the Sic.localspaces array has no more attributes.

Remember that an object is saved in the Sic.localspaces array if and only if an upper level variable is defined with the same name. Thus, at any time, the Python __main__ may contain objects from different levels. User does not have to care about the saving and unsaving mechanism: all is automatic and goes back as it was after procedure execution. He only has to know that he can access any lower level variable in the Sic.localspaces array.