This is a weird idea but it bounces around in my mind often enough that I thought I'd write it down.
What if ELSE and THEN Were PACK!-reactive?
Today they react only to antiform ghost and light null...with all other value states being considered "heavy enough to respond to".
But what if they required the weight of a PACK! to trigger, and anything that wasn't in a pack was deemed too light?
>> 10 then [print "This wouldn't print"]
== \~()~\ ; antiform (ghost!) "void"
>> pack [10] then [print "But this would print"]
But this would print
What's driven this to be on my mind is something about the irregularity of packs showing up only when you happen to produce a ghost or a light null. That makes them somewhat unpredictable to deal with.
So I've wondered... if PACK!s with only 1 element in them were made efficient (somehow?)... what would be the consequences of making taken branches always force things to be in a pack?
It might sound kind of bad because any ^META-assignment of a branch result would get a PACK!...not just two quirky cases.
But that's why I liked it: it was to force you to use SET-BLOCK! whenever you were going to ^META-assign the result of a branching structure.
^x: case [...]
if x [...] ; **error: oops, I got a weird pack
[^x]: case [...] ; okay, there. I fixed it.
This way your code doesn't suddenly get surprised the one day your branches make a null or a ghost and you get a pack.
The safety aspect is often on my mind, but in reality I don't know how to make this efficient... and problems from the current system don't happen very often in practice.
It's kind of a hazy idea. But since I have various places in the source that talk about it, I'm deleting those places and having a post to track any thoughts.
