addSlot complicating module handling

[Mark Thoma]

Hello,

why did you choose Slate to not group object slots statically (somehow like C
structs) in contrast to the current dynamic addSlot scheme? How do you want to
track which modules are safe to be unloaded and which might modify the slots of
other prototypes?
Does this level of dynamicity not introduce more problems - esp. with dynamic
module loading/unloading, but also speed - than it solves?
Has it proven to be so useful?

Regards,
Mark