You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The @windowed macro is something that doesn't sit very well with me. Its source code is just incredibly complicated. Why do we even need a macro for this, when it seems like a function could do the job just as well. Furthermore, why is the source code so complicated and why does it repeat so many things that are existing elsewhere in the library? It even repeats the definition of the RQA dictionary. It would be a very good idea to overhaul the windowed behavior as well, and make it a simpler function with shorter and clearer source code.
In Julia functions are first class objects. We can write something like:
windowed(x, f, args...; kwargs...)
and call f(args...) or so. I see no reason why we should be parsing expressions at all for a windowed functionality. In the end of the day, the logic is about indices accessing the recurrence matrix, and whether they overlap. We don't do any actual code manipulation, hence no reason for a macro.
The text was updated successfully, but these errors were encountered:
The
@windowed
macro is something that doesn't sit very well with me. Its source code is just incredibly complicated. Why do we even need a macro for this, when it seems like a function could do the job just as well. Furthermore, why is the source code so complicated and why does it repeat so many things that are existing elsewhere in the library? It even repeats the definition of the RQA dictionary. It would be a very good idea to overhaul the windowed behavior as well, and make it a simpler function with shorter and clearer source code.In Julia functions are first class objects. We can write something like:
and call
f(args...)
or so. I see no reason why we should be parsing expressions at all for a windowed functionality. In the end of the day, the logic is about indices accessing the recurrence matrix, and whether they overlap. We don't do any actual code manipulation, hence no reason for a macro.The text was updated successfully, but these errors were encountered: